.TH PLINK "1" "March 2020" "PLINK v2.00a3 64-bit (9 Apr 2020)" "User Commands" .SH NAME PLINK \- whole genome SNP analysis .SH DESCRIPTION PLINK v2.00a3 64\-bit (9 Apr 2020) www.cog\-genomics.org/plink/2.0/ (C) 2005\-2020 Shaun Purcell, Christopher Chang GNU General Public License v3 .PP In the command line flag definitions that follow, .IP * denote a required parameter, where the text between the .IP angle brackets describes its nature. .TP * ['square brackets + single\-quotes'] denotes an optional modifier. Use the .IP EXACT text in the quotes. .IP * [{bar|separated|braced|bracketed|values}] denotes a collection of mutually .TP exclusive optional modifiers (again, the exact text must be used). When .IP there are no outer square brackets, one of the choices must be selected. .IP * ['quoted_text='] denotes an optional modifier that .IP must begin with the quoted text, and be followed by a value with no whitespace in between. '|' may also be used here to indicate mutually exclusive options. .IP * [square brackets without quotes or braces] denote an optional parameter, .IP where the text between the brackets describes its nature. .IP * An ellipsis (...) indicates that you may enter multiple arguments of the .IP specified type. .IP * A "column set descriptor" is either .IP 1. a comma\-separated sequence of column set names; this is interpreted as .IP the full list of column sets to include. .IP 2. a comma\-separated sequence of column set names, all preceded by '+' or .IP \&'\-'; this is interpreted as a list of changes to the default. .IP plink2 [command flag(s)...] [other flag(s)...] plink2 \fB\-\-help\fR [flag name(s)...] .PP Most PLINK runs require exactly one main input fileset. The following flags are available for defining its form and location: .TP \fB\-\-pfile\fR ['vzs'] : Specify .pgen + .pvar[.zst] + .psam prefix. .TP \fB\-\-pgen\fR : Specify full name of .pgen/.bed file. .TP \fB\-\-pvar\fR : Specify full name of .pvar/.bim file. .TP \fB\-\-psam\fR : Specify full name of .psam/.fam file. .TP \fB\-\-bfile\fR ['vzs'] : Specify .bed + .bim[.zst] + .fam prefix. .HP \fB\-\-bpfile\fR ['vzs'] : Specify .pgen + .bim[.zst] + .fam prefix. .TP \fB\-\-keep\-autoconv\fR ['vzs'] : When importing non\-PLINK\-binary data, don't delete autogenerated fileset at end of run. .TP \fB\-\-no\-fid\fR : .fam file does not contain column 1 (family ID). .TP \fB\-\-no\-parents\fR : .fam file does not contain columns 3\-4 (parents). .TP \fB\-\-no\-sex\fR : .fam file does not contain column 5 (sex). .HP \fB\-\-vcf\fR ['dosage='] .HP \fB\-\-bcf\fR ['dosage='] : .IP Specify full name of .vcf{|.gz|.zst} or BCF2 file to import. * These can be used with \fB\-\-psam\fR/\-\-fam. * By default, dosage information is not imported. To import the GP field .IP (must be VCFv4.3\-style 0..1, one probability per possible genotype), add \&'dosage=GP' (or 'dosage=GP\-force', see below). To import Minimac3\-style DS+HDS phased dosage, add 'dosage=HDS'. 'dosage=DS' (or anything else for now) causes the named field to be interpreted as a Minimac3\-style dosage. Note that, in the dosage=GP case, PLINK 2 collapses the probabilities down to dosages; you cannot use PLINK 2 to losslessly convert VCF FORMAT:GP data to e.g. BGEN format. To make this more obvious, PLINK 2 now errors out when dosage=GP is used on a file with a FORMAT:DS header line and \fB\-\-import\-dosage\-certainty\fR wasn't specified, since dosage=DS extracts the same information more quickly in this situation. You can suppress this error with 'dosage=GP\-force'. In all of these cases, hardcalls are regenerated from scratch from the dosages. As a consequence, variants with no GT field can now be imported; they will be assumed to contain only diploid calls when HDS is also absent. .HP \fB\-\-data\fR ['gzs'] .HP \fB\-\-bgen\fR ['snpid\-chr'] .HP \fB\-\-gen\fR .HP \fB\-\-sample\fR : .TP Specify an Oxford\-format dataset to import. \fB\-\-data\fR specifies a .gen[.zst] .IP + .sample pair, while \fB\-\-bgen\fR specifies a BGEN v1.1+ file. * If a BGEN v1.2+ file contains sample IDs, it may be imported without a .IP companion .sample file. .IP * With 'snpid\-chr', chromosome codes are read from the 'SNP ID' field .IP instead of the usual chromosome field. .IP * The following REF/ALT modes are supported: .IP \&'ref\-first': The first allele for each variant is REF. \&'ref\-last': The last allele for each variant is REF. \&'ref\-unknown': The last allele for each variant is treated as .IP provisional\-REF. .HP \fB\-\-haps\fR [{ref\-first | ref\-last}] .HP \fB\-\-legend\fR : .IP Specify .haps [+ .legend] file(s) to import. * When \fB\-\-legend\fR is specified, it's assumed that the \fB\-\-haps\fR file doesn't .IP contain header columns. .TP * On chrX, the second male column may contain dummy '\-' entries. (However, .IP PLINK 2 currently cannot handle omitted male columns.) .IP * If not used with \fB\-\-sample\fR, new sample IDs are of the form 'per#/per#'. .TP \fB\-\-map\fR : Specify full name of .map file. .HP \fB\-\-import\-dosage\fR ['noheader'] ['id\-delim='] .IP ['skip0='] ['skip1='] ['skip2='] ['dose1'] ['format='] [{ref\-first | ref\-last}] ['single\-chr='] ['chr\-col\-num='<#>] ['pos\-col\-num='<#>] : .IP Specify PLINK 1.x\-style dosage file to import. * You must also specify a companion .psam/.fam file. * By default, PLINK assumes that the file contains a header line, which has .IP \&'SNP' in (1\-based) column i+1, 'A1' in column i+j+2, 'A2' in column i+j+3, and sample FID/IIDs starting from column i+j+k+4. (i/j/k are normally zero, but can be changed with 'skip0', 'skip1', and 'skip2' respectively. FID/IID are normally assumed to be separate tokens, but if they're merged into a single token you can specify the delimiter with \&'id\-delim='.) If such a header line is not present, use the 'noheader' modifier; samples will then be assumed to appear in the same order as they do in the .psam/.fam file. .TP * You may specify a companion .map file. If you do not, .IP * 'single\-chr=' can be used to specify that all variants are on the named .TP chromosome. Otherwise, you can use 'chr\-col\-num=' to read chromosome .IP codes from the given (1\-based) column number. .IP * 'pos\-col\-num=' causes bp coordinates to be read from the given column .IP number. .IP * The 'format=' modifier lets you specify the number of values used to .TP represent each dosage. \&'format=1' normally indicates a single 0..2 A1 .TP expected count; 'dose1' modifies this to a 0..1 frequency. \&'format=2' .IP indicates a 0..1 homozygous A1 likelihood followed by a 0..1 het likelihood. 'format=3' indicates 0..1 hom A1, 0..1 het, 0..1 hom A2. \&'format=infer' (the default) infers the format from the number of columns in the first nonheader line. .HP \fB\-\-dummy\fR [missing dosage freq] [missing pheno freq] .IP [{acgt | 1234 | 12}] ['pheno\-ct='] ['scalar\-pheno'] ['dosage\-freq='] .IP This generates a fake input dataset with the specified number of samples and SNPs. * By default, the missing dosage and phenotype frequencies are zero. .IP These can be changed by providing 3rd and 4th numeric arguments. .IP * By default, allele codes are As and Bs; this can be changed with the .IP \&'acgt', '1234', or '12' modifier. .TP * By default, one binary phenotype is generated. \&'pheno\-ct=' can be used .IP to change the number of phenotypes, and 'scalar\-pheno' causes these phenotypes to be normally distributed scalars. .TP * By default, all (nonmissing) dosages are in {0,1,2}. To make some of .TP them take on decimal values, use 'dosage\-freq='. (These dosages are .IP affected by \fB\-\-hard\-call\-threshold\fR and \fB\-\-dosage\-erase\-threshold\fR.) .TP \fB\-\-fa\fR : Specify full name of reference FASTA file. .PP Output files have names of the form 'plink2.' by default. You can change the 'plink2' prefix with .TP \fB\-\-out\fR : Specify prefix for output files. .PP Most runs also require at least one of the following commands: .HP \fB\-\-rm\-dup\fR [mode] ['list'] .IP Remove all but one instance of each duplicate\-ID variant (ignoring the missing ID), and (with the 'list' modifier) write a list of duplicated IDs to .rmdup.list. The following modes of operation are supported: * 'error' (default) causes this to error out when there's a genotype data .TP or other mismatch between the records. A list of affected IDs is written .IP to .rmdup.mismatch. .IP * 'retain\-mismatch' causes all instances of a duplicate\-ID variant to be .IP retained when there's a genotype data or variant info mismatch; otherwise one instance is kept. The .rmdup.mismatch file is also written. .IP * 'exclude\-mismatch' removes all instances of duplicate\-ID mismatched .IP variants instead. .IP * 'exclude\-all' causes all instances of duplicate\-ID variants to be .IP removed, even when the actual records are identical. .IP * 'force\-first' causes only the first instance of duplicate\-ID variants to .IP be kept, under all circumstances. .HP \fB\-\-make\-pgen\fR ['vzs'] ['format='] ['trim\-alts'] ['erase\-phase'] .IP ['erase\-dosage'] ['pvar\-cols='] ['psam\-cols='] .HP \fB\-\-make\-bpgen\fR ['vzs'] ['format='] ['trim\-alts'] ['erase\-phase'] .IP ['erase\-dosage'] .HP \fB\-\-make\-bed\fR ['vzs'] ['trim\-alts'] .IP Create a new PLINK 2 binary fileset (\fB\-\-make\-pgen\fR = .pgen + .pvar[.zst] + \&.psam, \fB\-\-make\-bpgen\fR = .pgen + .bim[.zst] + .fam). * Unlike the automatic text\-to\-binary converters (which only heed .IP chromosome filters), this supports all of PLINK's filtering flags. .IP * The 'vzs' modifier causes the variant file (.pvar/.bim) to be .IP Zstd\-compressed. .IP * The 'format' modifier requests an uncompressed fixed\-variant\-width .pgen .TP file. (These do not directly support multiallelic variants.) The .IP following format code is currently supported: .IP 2: just like .bed, except with an extended (12\-byte instead of 3\-byte) .IP header containing variant/sample counts, and rotated genotype codes (00 = hom ref, 01 = het, 10 = hom alt, 11 = missing). .IP * The 'erase\-phase' and 'erase\-dosage' modifiers prevent phase and dosage .IP information from being written to the new .pgen. .IP * The first five columns of a .pvar file are always #CHROM/POS/ID/REF/ALT. .IP Supported optional .pvar column sets are: .IP xheader: All ## header lines (yeah, this is technically not a column), .IP except for possibly FILTER/INFO definitions when those column(s) have been removed. Without this, only the #CHROM header line is kept. .IP vcfheader: xheader, with additions to make it a valid VCF header. maybequal: QUAL. Omitted if all remaining values are missing. qual: Force QUAL column to be written even when empty. maybefilter: FILTER. Omitted if all remaining values are missing. filter: Force FILTER column to be written even when empty. maybeinfo: INFO. Omitted if all remaining values are missing, or if .IP INFO:PR is the only subfield. .IP info: Force INFO column to be written. maybecm: Centimorgan coordinate. Omitted if all remaining values = 0. cm: Force CM column to be written even when empty. .IP The default is xheader,maybequal,maybefilter,maybeinfo,maybecm. .IP * Supported column sets for the .psam file are: .TP maybefid: Family ID, '0' = missing. Omitted if all values missing. .IP fid: Force FID column to be written even when empty. maybesid: Source ID, '0' = missing. Omitted if all values missing. sid: Force SID column to be written even when empty. maybeparents: Father and mother IIDs. Omitted if all values missing. parents: Force PAT and MAT columns to be written even when empty. sex: '1' = male, '2' = female, 'NA' = missing. pheno1: First active phenotype. If none, all column entries are set to .IP the \fB\-\-output\-missing\-phenotype\fR string. .TP phenos: All active phenotypes, if any. (Can be combined with pheno1 to .IP force at least one phenotype column to be written.) .IP The default is maybefid,maybesid,maybeparents,sex,phenos. .HP \fB\-\-make\-just\-pvar\fR ['zs'] ['cols='] .HP \fB\-\-make\-just\-psam\fR ['cols='] .HP \fB\-\-make\-just\-bim\fR ['zs'] .HP \fB\-\-make\-just\-fam\fR .IP Variants of \fB\-\-make\-pgen\fR/\-\-make\-bed which only write a new variant or sample file. These don't always require an input genotype file. USE THESE CAUTIOUSLY. It is very easy to desynchronize your binary genotype data and your sample/variant indexes if you use these commands improperly. If you have any doubt, stick with \fB\-\-make\-[b]pgen\fR/\-\-make\-bed. .HP \fB\-\-export\fR [{01 | 12}] ['bgz'] ['id\-delim='] .IP ['id\-paste='] ['include\-alt'] ['omit\-nonmale\-y'] ['spaces'] ['vcf\-dosage='] ['ref\-first'] ['bits='<#>] .TP Create a new fileset with all filters applied. The following output .IP formats are supported: (actually, only A, AD, A\-transpose, bcf, bgen\-1.x, haps, hapslegend, ind\-major\-bed, oxford, and vcf are implemented for now) * '23': 23andMe 4\-column format. This can only be used on a single .IP sample's data (\fB\-\-keep\fR may be handy), and does not support multicharacter allele codes. .IP * 'A': Sample\-major additive (0/1/2) coding, suitable for loading from R. .IP If you need uncounted alleles to be named in the header line, add the 'include\-alt' modifier. .IP * 'AD': Sample\-major additive (0/1/2) + dominant (het=1/hom=0) coding. .IP Also supports 'include\-alt'. .IP * 'A\-transpose': Variant\-major 0/1/2. * 'beagle': Unphased per\-autosome .dat and .map files, readable by early .IP BEAGLE versions. .IP * 'beagle\-nomap': Single .beagle.dat file. * 'bgen\-1.x': Oxford\-format .bgen + .sample. For v1.2/v1.3, sample .IP identifiers are stored in the .bgen (with id\-delim and id\-paste settings applied), and default precision is 16\-bit (use the 'bits' modifier to reduce this). .IP * 'bimbam': Regular BIMBAM format. * 'bimbam\-1chr': BIMBAM format, with a two\-column .pos.txt file. Does not .IP support multiple chromosomes. .IP * 'fastphase': Per\-chromosome fastPHASE files, with .IP \&.chr\-.phase.inp filename extensions. .TP * 'fastphase\-1chr': Single .phase.inp file. Does not support .IP multiple chromosomes. .TP * 'haps', 'hapslegend': Oxford\-format .haps + .sample[ + .legend]. All .TP data must be biallelic and phased. When the 'bgz' .TP modifier is present, the .haps file is block\-gzipped. .IP * 'HV': Per\-chromosome Haploview files, with .chr\-{.ped,.info} .IP filename extensions. .TP * 'HV\-1chr': Single Haploview .ped + .info file pair. Does not support .IP multiple chromosomes. .IP * 'ind\-major\-bed': PLINK 1 sample\-major .bed (+ .bim + .fam). * 'lgen': PLINK 1 long\-format (.lgen + .fam + .map), loadable with \fB\-\-lfile\fR. * 'lgen\-ref': .lgen + .fam + .map + .ref, loadable with \fB\-\-lfile\fR + .IP \fB\-\-reference\fR. .TP * 'list': Single genotype\-based list, up to 4 lines per variant. To omit .IP nonmale genotypes on the Y chromosome, add the 'omit\-nonmale\-y' modifier. .IP * 'rlist': .rlist + .fam + .map fileset, where the .rlist file is a .IP genotype\-based list which omits the most common genotype for each variant. Also supports 'omit\-nonmale\-y'. .TP * 'oxford': Oxford\-format .gen + .sample. When the 'bgz' modifier is .IP present, the .gen file is block\-gzipped. .IP * 'ped': PLINK 1 sample\-major (.ped + .map), loadable with \fB\-\-file\fR. * 'compound\-genotypes': Same as 'ped', except that the space between each .IP pair of same\-variant allele codes is removed. .IP * 'structure': Structure\-format. * 'transpose': PLINK 1 variant\-major (.tped + .tfam), loadable with .IP \fB\-\-tfile\fR. .TP * 'vcf', : VCF (default version 4.3). If PAR1 and PAR2 are present, .TP \&'vcf\-4.2', they are automatically merged with chrX, with proper .TP \&'bcf', handling of chromosome codes and male ploidy. .TP \&'bcf\-4.2' When the 'bgz' modifier is present, the VCF file is block\-gzipped. (This always happens with BCF output.) The 'id\-paste' modifier controls which .psam columns are used to construct sample IDs (choices are maybefid, fid, iid, maybesid, and sid; default is maybefid,iid,maybesid), while the 'id\-delim' modifier sets the character between the ID pieces (default '_'). Genotypes are always exported. If you want to export a sites\-only VCF instead, see \fB\-\-make\-pgen\fR/\-\-make\-just\-pvar's \&'vcfheader' column set. Dosages are not exported unless the 'vcf\-dosage=' modifier is present. The following five dosage export modes are supported: \&'GP': genotype posterior probabilities (v4.3 only). \&'DS': Minimac3\-style dosages, omitted for hardcalls. \&'DS\-force': Minimac3\-style dosages, never omit. \&'HDS': Minimac3\-style phased dosages, omitted for hardcalls .TP and unphased calls. Also includes 'DS' output. .IP \&'HDS\-force': Always report DS and HDS. .IP In addition, * The '12' modifier causes alt1 alleles to be coded as '1' and ref alleles .IP to be coded as '2', while '01' maps alt1 \-> 0 and ref \-> 1. .IP * The 'spaces' modifier makes the output space\-delimited instead of .IP tab\-delimited, whenever both are permitted. .IP * For biallelic formats where it's unspecified whether the reference/major .IP allele should appear first or second, \fB\-\-export\fR defaults to second for compatibility with PLINK 1.9. Use 'ref\-first' to change this. (Note that this doesn't apply to the 'A', 'AD', and 'A\-transpose' formats; use \fB\-\-export\-allele\fR to control which alleles are counted there.) .HP \fB\-\-freq\fR ['zs'] ['counts'] ['cols='] ['bins\-only'] .IP ['refbins=' | 'refbins\-file='] ['alt1bins=' | 'alt1bins\-file='] .TP Empirical allele frequency report. By default, only founders are .TP considered. Dosages are taken into account (e.g. heterozygous haploid .IP calls count as 0.5). Supported column sets are: .IP chrom: Chromosome ID. pos: Base\-pair coordinate. (ID is always present, and positioned here.) ref: Reference allele. alt1: Alternate allele 1. alt: All alternate alleles, comma\-separated. reffreq: Reference allele frequency/dosage. alt1freq: Alt1 frequency/dosage. altfreq: Comma\-separated frequencies/dosages for all alternate alleles. freq: Similar to altfreq, except ref is also included at the start. eq: Comma\-separated = for all present alleles. (If no .IP alleles are present, the column contains a single '.'.) .IP eqz: Same as eq, except zero\-counts are included. alteq/alteqz: Same as eq/eqz, except reference allele is omitted. numeq: 0=,1=, etc. Zero\-counts are omitted. altnumeq: Same as numeq, except reference allele is omitted. machr2: Unphased MaCH imputation quality metric. minimac3r2: Phased Minimac3 imputation quality. nobs: Number of allele observations. .IP The default is chrom,ref,alt,altfreq,nobs. Additional .afreq.{ref,alt1}.bins (or .acount.{ref,alt1}.bins with \&'counts') file(s) are generated when 'refbins='/'refbins\-file=' or \&'alt1bins='/'alt1bins\-file=' is present; these report the total number of frequencies or counts in each left\-closed, right\-open interval. (If you only want these histogram(s), and not the main report, add 'bins\-only'.) .HP \fB\-\-geno\-counts\fR ['zs'] ['cols='] .IP Variant\-based hardcall genotype count report (considering both alleles simultaneously in the diploid case). Nonfounders are now included; use \fB\-\-keep\-founders\fR if this is a problem. Heterozygous haploid calls are treated as missing. Supported column sets are: .IP chrom: Chromosome ID. pos: Base\-pair coordinate. (ID is always present, and positioned here.) ref: Reference allele. alt1: Alternate allele 1. alt: All alternate alleles, comma\-separated. homref: Homozygous\-ref count. refalt1: Heterozygous ref\-alt1 count. refalt: Comma\-separated het ref\-altx counts. homalt1: Homozygous\-alt1 count. altxy: Comma\-separated altx\-alty counts, in (1/1)\-(1/2)\-(2/2)\-(1/3)\-... .IP order. .IP xy: Similar to altxy, except the reference allele is treated as alt0, .IP and the sequence starts (0/0)\-(0/1)\-(1/1)\-(0/2)\-... .IP hapref: Haploid\-ref count. hapalt1: Haploid\-alt1 count. hapalt: Comma\-separated haploid\-altx counts. hap: Similar to hapalts, except ref is also included at the start. numeq: 0/0=,0/1=,1/1=,...,0= .TP etc. Zero\-counts are omitted. (If all genotypes are missing, the .IP column contains a single '.'.) .IP missing: Number of missing genotypes. nobs: Number of (nonmissing) genotype observations. .IP The default is chrom,ref,alt,homref,refalt,altxy,hapref,hapalt,missing. .HP \fB\-\-sample\-counts\fR ['zs'] ['cols='] .IP Sample\-based hardcall genotype count report. * Unknown\-sex samples are treated as female. * Heterozygous haploid calls (MT included) are treated as missing. * As with other PLINK 2 commands, SNPs that have not been left\-normalized .TP are counted as non\-SNP non\-symbolic. (Use e.g. \fB\-\-normalize\fR when that's a .IP problem.) .IP * Supported column sets are: .IP maybefid: FID, if that column was present in the input. fid: Force FID column to be written even when absent in the input. (IID is always present, and positioned here.) maybesid: SID, if that column was present in the input. sid: Force SID column to be written even when absent in the input. sex: '1' = male, '2' = female, 'NA' = missing. hom: Homozygous genotype count. homref: Homozygous\-ref genotype count. homalt: Homozygous\-alt genotype count. homaltsnp: Homozygous\-alt SNP count. het: Heterozygous genotype count. refalt: Heterozygous ref\-altx genotype count. het2alt: Heterozygous altx\-alty genotype count. hetsnp: Heterozygous SNP count. dipts: Diploid SNP transition count. ts: SNP transition count (excluding chrY for females). diptv: Diploid SNP transversion count. tv: SNP transversion count. dipnonsnpsymb: Diploid non\-SNP, non\-symbolic count. nonsnpsymb: Non\-SNP, non\-symbolic count. symbolic: Symbolic variant count. nonsnp: Non\-SNP count. dipsingle: Number of singletons relative to this dataset, across just .TP diploid calls. (Note that if the ALT allele in a chrX .IP biallelic variant appears in exactly one female and one male, that counts as a singleton for just the female.) .IP single: Number of singletons relative to this dataset. haprefwfemaley: Haploid\-ref count, counting chrY for everyone. hapref: Haploid\-ref count, excluding chrY for females. hapaltwfemaley: Haploid\-alt count, counting chrY for everyone. hapalt: Haploid\-alt count, excluding chrY for females. missingwfemaley: Missing call count, counting chrY for everyone. missing: Missing call count, excluding chrY for females. .IP The default is maybefid,maybesid,homref,homaltsnp,hetsnp,dipts,diptv, dipnonsnpsymb,dipsingle,haprefwfemaley,hapaltwfemaley,missingwfemaley. .IP * The 'hetsnp', 'dipts'/'ts'/'diptv'/'tv', 'dipnonsnpsymb'/'nonsnpsymb', .IP \&'symbolic', and 'nonsnp' columns count each ALT allele in a heterozygous altx\-alty call separately, since they can be of different subtypes. (I.e. if they are of the same subtype, the corresponding count is incremented by 2.) As a consequence, these columns are unaffected by variant split/join. .HP \fB\-\-missing\fR ['zs'] [{sample\-only | variant\-only}] .IP ['scols='] ['vcols='] .IP Generate sample\- and variant\-based missing data reports (or just one report if 'sample\-only'/'variant\-only' is specified). As of alpha 2, mixed MT hardcalls appear in the heterozygous haploid stats. Supported column sets in the sample\-based report are: .IP maybefid: FID, if that column was present in the input. fid: Force FID column to be written even when absent in the input. (IID is always present, and positioned here.) maybesid: SID, if that column was present in the input. sid: Force SID column to be written even when absent in the input. misspheno1: First active phenotype missing (Y/N)? Always 'Y' if no .IP phenotypes are loaded. .TP missphenos: A Y/N column for each loaded phenotype. (Can be combined .IP with misspheno1 to force at least one such column.) .IP nmissdosage: Number of missing dosages. nmiss: Number of missing hardcalls, not counting het haploids. nmisshh: Number of missing hardcalls, counting het haploids. hethap: Number of heterozygous haploid hardcalls. nobs: Denominator (male count on chrY, otherwise total sample count). fmissdosage: Missing dosage rate. fmiss: Missing hardcall rate, not counting het haploids. fmisshh: Missing hardcall rate, counting het haploids. .IP The default is maybefid,maybesid,missphenos,nmiss,nobs,fmiss. Supported column sets in the variant\-based report are: .IP chrom: Chromosome ID. pos: Base\-pair coordinate. (ID is always present, and positioned here.) ref: Reference allele. alt1: Alternate allele 1. alt: All alternate alleles, comma\-separated. nmissdosage: Number of missing dosages. nmiss: Number of missing hardcalls, not counting het haploids. nmisshh: Number of missing hardcalls, counting het haploids. hethap: Number of heterozygous haploid calls. nobs: Number of potentially valid calls. fmissdosage: Missing dosage rate. fmiss: Missing hardcall rate, not counting het haploids. fmisshh: Missing hardcall rate, counting het haploids. fhethap: Heterozygous haploid rate. .IP The default is chrom,nmiss,nobs,fmiss. .HP \fB\-\-hardy\fR ['zs'] ['midp'] ['redundant'] ['cols='] .IP Hardy\-Weinberg exact test p\-value report(s). * By default, only founders are considered; change this with \fB\-\-nonfounders\fR. * chrX is now omitted from the main .hardy report. Instead, .IP (if present) it gets its own .hardy.x report based on the method described in Graffelman J, Weir BS (2016) Hardy\-Weinberg equilibrium and the X chromosome. .IP * For variants with k alleles where k>2, k separate 'biallelic' tests are .TP performed, each reported on its own line. However, biallelic variants .IP are normally reported on a single line, since the counts/frequencies would be mirror\-images and the p\-values would be the same. You can add the 'redundant' modifier to force biallelic variant results to be reported on two lines for parsing convenience. .IP * There is currently no special handling of case/control phenotypes. Supported column sets are: .IP chrom: Chromosome ID. pos: Base\-pair coordinate. (ID is always present, and positioned here.) ref: Reference allele. alt1: Alternate allele 1. alt: All alternate alleles, comma\-separated. (A1 is always present, and positioned here.) ax: Non\-A1 allele(s), comma\-separated. gcounts: Hom\-A1 count, total number of het\-A1 calls, and total number of .TP nonmissing calls with no copies of A1. On chrX, these are .IP followed by male A1 and male non\-A1 counts. .IP gcount1col: gcounts values in a single comma\-separated column. hetfreq: Observed and expected het\-A1 frequencies. sexaf: Female and male A1 observed allele frequencies (chrX only). femalep: Female\-only p/midp\-value (chrX only). p: Hardy\-Weinberg equilibrium exact test p/midp\-value. .IP The default is chrom,ax,gcounts,hetfreq,sexaf,p. .HP \fB\-\-indep\-pairwise\fR ['kb'] [step size (variant ct)] .IP .IP Generate a list of variants in approximate linkage equilibrium. * For multiallelic variants, major allele counts are used in the r^2 .IP computation. .IP * With the 'kb' modifier, the window size is in kilobase instead of variant .TP count units. (Pre\-'kb' space is optional, i.e. .IP "\-\-indep\-pairwise 500 kb 0.5" and "\-\-indep\-pairwise 500kb 0.5" have the same effect.) .IP * The step size now defaults to 1 if it's unspecified, and *must* be 1 if .IP the window is in kilobase units. .IP * Note that you need to rerun PLINK using \fB\-\-extract\fR or \fB\-\-exclude\fR on the .IP \&.prune.in/.prune.out file to apply the list to another computation... and as with other applications of \fB\-\-extract\fR/\-\-exclude, duplicate variant IDs are a problem. \fB\-\-indep\-pairwise\fR still runs to completion for now when duplicate variant IDs are present, but that will become an error in alpha 3. .HP \fB\-\-ld\fR ['dosage'] ['hwe\-midp'] .IP This displays diplotype frequencies, r^2, and D' for a single pair of variants. * For multiallelic variants, major allele counts/dosages are used. * Phase information is used when both variants are on the same chromosome. * When there is at least one sample with unphased het calls for both .IP variants, diplotype frequencies are estimated using the Hill equation. If there are multiple biologically possible local maxima, all are displayed, along with HWE exact test statistics. .TP * By default, only hardcalls are considered. Add the 'dosage' modifier if .TP you want dosages to be taken into account. (In the diploid case, an .IP unphased dosage of x is interpreted as P(0/0) = 1 \- x, P(0/1) = x when x is in 0..1.) .HP \fB\-\-sample\-diff\fR ['id\-delim='] ['dosage' | 'dosage='] .IP ['include\-missing'] [{pairwise | counts\-only}] ['fname\-id\-delim='] ['zs'] ['cols='] ['counts\-cols='] {base= | ids=} [other sample ID(s)...] .HP \fB\-\-sample\-diff\fR ['id\-delim='] ['dosage' | 'dosage='] .IP ['include\-missing'] [{pairwise | counts\-only}] ['fname\-id\-delim='] ['zs'] ['cols='] ['counts\-cols='] file= .IP (alias: \fB\-\-sdiff\fR) Report discordances and discordance\-counts between pairs of samples. If chrX or chrY is present, sex must be defined and consistent. * There are three ways to specify which sample pairs to compare. To .IP compare a single baseline sample against some others, start the (space\-delimited) sample ID list with 'base='. To perform an all\-vs.\-all comparison, start it with 'ids=' instead. To compare sample pairs listed in a file, use 'file='. Note that 'base='/'ids='/'file=' must be positioned after all modifiers. .IP * Sample IDs are interpreted as if they were in a VCF header line, with .IP \&'id\-delim=' having the usual effect. .TP * By default, comparisons are based on hardcalls. Use 'dosage' to compare .IP dosages instead; you can combine this with a tolerance in [0, 0.5). .IP * By default, if one genotype is missing and the other isn't, that doesn't .IP count as a difference; this can be changed with 'include\-missing'. .IP * By default, a single main report is written to .TP [.].sdiff. To write separate pairwise .IP ...sdiff reports for each compared ID pair, add the 'pairwise' modifier. To omit the main report, add the 'counts\-only' modifier. (Note that, if you're only interested in nonmissing autosomal biallelic hardcalls, \fB\-\-make\-king\-table\fR provides a more efficient way to compute just counts.) .IP * By default, if an output filename has a multipart sample ID, the parts .IP will be delimited by '_'; use 'fname\-id\-delim=' to change this. .IP Supported main\-report column sets are: .IP chrom: Chromosome ID. pos: Base\-pair coordinate. (Variant ID is always present, and positioned here.) ref: Reference allele. alt: All alternate alleles, comma\-separated. maybefid: FID1/FID2, if that column was in the input. Requires 'id'. fid: Force FID1/FID2 even when FID was absent in the input. id: IID1/IID2. maybesid: SID1/SID2, if that column was in the input. Requires 'id'. sid: Force SID1/SID2 even when SID was absent in the input. geno: Unphased GT or DS for the two samples. .IP The default is usually chrom,pos,ref,alt,maybefid,id,maybesid,geno; the sample IDs are removed from the default in 'pairwise' mode. Supported discordance\-count\-summary column sets are: .IP maybefid: FID1/FID2, if that column was in the input. fid: Force FID1/FID2 even when FID was absent in the input. (IID1/IID2 are always present.) maybesid: SID1/SID2, if that column was in the input. sid: Force SID1/SID2 even when SID was absent in the input. nobs: Number of variants considered. This includes variants where one or .IP both variants are missing iff 'include\-missing' was specified. .IP nobsibs: ibs0+ibs1+ibs2. ibs0: Number of diploid variants with no common hardcall alleles. ibs1: Number of diploid variants with exactly 1 common hardcall allele. ibs2: Number of diploid variants with both hardcall alleles matching. halfmiss: Number of variants with exactly 1 missing genotype/dosage. .IP Ignored without 'include\-missing'. .IP diff: Total number of differences. .IP The default is maybefid,maybesid,nobs,halfmiss,diff. .HP \fB\-\-make\-king\fR [{square | square0 | triangle}] [{zs | bin | bin4}] .IP KING\-robust kinship estimator, described by Manichaikul A, Mychaleckyj JC, Rich SS, Daly K, Sale M, Chen WM (2010) Robust relationship inference in genome\-wide association studies. By default, this writes a lower\-triangular tab\-delimited table of kinship coefficients to .king, and a list of the corresponding sample IDs to .king.id. The first row of the .king file contains a single kinship coefficient, the second row has the and kinship values in that order, etc. * Only autosomes are currently considered. * Pedigree information is currently ignored; the between\-family estimator .IP is used for all pairs. .IP * For multiallelic variants, REF allele counts are used. * If the 'square' or 'square0' modifier is present, a square matrix is .IP written instead; 'square0' fills the upper right triangle with zeroes. .IP * If the 'zs' modifier is present, the .king file is Zstd\-compressed. * If the 'bin' modifier is present, a binary (square) matrix of .IP double\-precision floating point values, suitable for loading from R, is instead written to .king.bin. ('bin4' specifies single\-precision numbers instead.) This can be combined with 'square0' if you still want the upper right zeroed out, or 'triangle' if you don't want to pad the upper right at all. .IP * The computation can be subdivided with \fB\-\-parallel\fR. .HP \fB\-\-make\-king\-table\fR ['zs'] ['counts'] ['rel\-check'] ['cols='] .IP Similar to \fB\-\-make\-king\fR, except results are reported in KING's original \&.kin0 text table format (with minor changes, e.g. row order is more friendly to incremental addition of samples), \fB\-\-king\-table\-filter\fR can be used to restrict the report to high kinship values, and the 'rel\-check' modifier can be used to restrict to same\-FID pairs. Supported column sets are: .TP maybefid: FID1/FID2, if that column was in the input. Requires 'id'. .IP fid: Force FID1/FID2 even when FID was absent in the input. id: IID1/IID2. maybesid: SID1/SID2, if that column was in the input. Requires 'id'. sid: Force SID1/SID2 even when SID was absent in the input. nsnp: Number of variants considered (autosomal, neither call missing). hethet: Proportion/count of considered call pairs which are het\-het. ibs0: Proportion/count of considered call pairs which are opposite homs. ibs1: HET1_HOM2 and HET2_HOM1 proportions/counts. kinship: KING\-robust between\-family kinship estimator. .IP The default is maybefid,id,maybesid,nsnp,hethet,ibs0,kinship. hethet/ibs0/ibs1 values are proportions unless the 'counts' modifier is present. If id is omitted, a .kin0.id file is also written. .HP \fB\-\-make\-rel\fR ['cov'] ['meanimpute'] [{square | square0 | triangle}] .IP [{zs | bin | bin4}] .IP Write a lower\-triangular variance\-standardized relationship matrix to .rel, and corresponding IDs to .rel.id. * This computation assumes that variants do not have very low MAF, or .IP deviate greatly from Hardy\-Weinberg equilibrium. .IP * Also, it's usually best to perform this calculation on a variant set in .IP approximate linkage equilibrium. .IP * The 'cov' modifier replaces the variance\-standardization step with basic .IP mean\-centering, causing a covariance matrix to be calculated instead. .IP * The computation can be subdivided with \fB\-\-parallel\fR. .HP \fB\-\-make\-grm\-list\fR ['cov'] ['meanimpute'] ['zs'] [{id\-header | iid\-only}] .HP \fB\-\-make\-grm\-bin\fR ['cov'] ['meanimpute'] [{id\-header | iid\-only}] .HP \fB\-\-make\-grm\-list\fR causes the relationships to be written to GCTA's original .IP list format, which describes one pair per line, while \fB\-\-make\-grm\-bin\fR writes them in GCTA 1.1+'s single\-precision triangular binary format. Note that these formats explicitly report the number of valid observations (where neither sample has a missing call) for each pair, which is useful input for some scripts. .HP \fB\-\-pca\fR [count] [{approx | meanimpute}] ['scols='] .HP \fB\-\-pca\fR [{allele\-wts | biallelic\-var\-wts}] [count] [{approx | meanimpute}] .IP ['vzs'] ['scols='] ['vcols='] .IP Extracts top principal components from the variance\-standardized relationship matrix. * It is usually best to perform this calculation on a variant set in .IP approximate linkage equilibrium, with no very\-low\-MAF variants. .IP * By default, 10 PCs are extracted; you can adjust this by passing a .TP numeric argument. (Note that 10 is lower than the PLINK 1.9 default of .IP 20; this is due to the randomized algorithm's memory footprint growing quadratically w.r.t. the PC count.) .IP * The 'approx' modifier causes the standard deterministic computation to be .IP replaced with the randomized algorithm originally implemented for Galinsky KJ, Bhatia G, Loh PR, Georgiev S, Mukherjee S, Patterson NJ, Price AL (2016) Fast Principal\-Component Analysis Reveals Convergent Evolution of ADH1B in Europe and East Asia. This can be a good idea when you have >5k samples, and is almost required with >50k. .IP * The randomized algorithm always uses mean imputation for missing genotype .TP calls. For comparison purposes, you can use the 'meanimpute' modifier to .IP request this behavior for the standard computation. .IP * 'scols=' can be used to customize how sample IDs appear in the .eigenvec .TP file. (maybefid, fid, maybesid, and sid supported; default is .IP maybefid,maybesid.) .IP * The 'allele\-wts' modifier requests an additional one\-line\-per\-allele .IP \&.eigenvec.allele file with PCs expressed as allele weights instead of sample weights. When it's present, 'vzs' causes the .eigenvec.allele file to be Zstd\-compressed. \&'vcols=' can be used to customize the report columns; supported column sets are: .IP chrom: Chromosome ID. pos: Base\-pair coordinate. (ID is always present, and positioned here.) ref: Reference allele. alt1: Alternate allele 1. alt: All alternate alleles, comma\-separated. (A1 is always present, and positioned here.) ax: Non\-A1 alleles, comma\-separated. (PCs are always present, and positioned here.) .IP Default is chrom,ref,alt. .IP * For datasets with no multiallelic variants, the 'biallelic\-var\-wts' .IP modifier requests the old .eigenvec.var format, which only reports weights for major alleles. (These weights are 2x the corresponding \&.eigenvec.allele weights.) Supported column sets are: .IP chrom: Chromosome ID. pos: Base\-pair coordinate. (ID is always present, and positioned here.) ref: Reference allele. alt1: Alternate allele 1. alt: All alternate alleles, comma\-separated. maj: Major allele. nonmaj: Minor allele. (PCs are always present, and positioned here. Signs are w.r.t. the major, not necessarily reference, allele.) .IP Default is chrom,maj,nonmaj. .HP \fB\-\-king\-cutoff\fR [.king.bin + .king.id fileset prefix] .IP Exclude one member of each pair of samples with KING\-robust kinship greater than the given threshold. Remaining/excluded sample IDs are written to .king.cutoff.in.id + .king.cutoff.out.id. If present, the .king.bin file must be triangular (either precision is ok). .HP \fB\-\-write\-covar\fR ['cols='] .IP If covariates are defined, an updated version (with all filters applied) is automatically written to .cov whenever \fB\-\-make\-pgen\fR, \fB\-\-make\-just\-psam\fR, \fB\-\-export\fR, or a similar command is present. However, if you do not wish to simultaneously generate a new sample file, you can use \fB\-\-write\-covar\fR to just produce a pruned covariate file. Supported column sets are: .IP maybefid: FID, if that column was in the input. fid: Force FID column to be written even when absent in the input. maybesid: SID, if that column was in the input. sid: Force SID column to be written even when absent in the input. maybeparents: Father/mother IIDs ('0' = missing), if columns in input. parents: Force PAT/MAT columns to be written even when absent in input. sex: '1' = male, '2' = female, 'NA' = missing. pheno1: First active phenotype. If none, all column entries are set to .IP the \fB\-\-output\-missing\-phenotype\fR string. .TP phenos: All active phenotypes, if any. (Can be combined with pheno1 to .IP force at least one phenotype column to be written.) .IP (Covariates are always present, and positioned here.) .IP The default is maybefid,maybesid. .HP \fB\-\-write\-samples\fR .IP Report IDs of all samples which pass your filters/inclusion thresholds. .HP \fB\-\-write\-snplist\fR ['zs'] ['allow\-dups'] .TP List all variants which pass your filters/inclusion thresholds. Unless the .IP \&'allow\-dups' modifier is provided, this now errors out when duplicate variant ID(s) remain. .HP \fB\-\-glm\fR ['zs'] ['omit\-ref'] [{sex | no\-x\-sex}] ['log10'] ['pheno\-ids'] .IP [{genotypic | hethom | dominant | recessive}] ['interaction'] ['skip'] ['hide\-covar'] [{no\-firth | firth\-fallback | firth}] ['intercept'] ['cols='] ['local\-covar='] ['local\-psam='] ['local\-pos\-cols=' | 'local\-pvar='] ['local\-haps'] ['local\-omit\-last' | 'local\-cats[0]='] ['allow\-no\-covars'] .IP Basic association analysis on quantitative and/or case/control phenotypes. For each variant, a linear (for quantitative traits) or logistic (for case/control) regression is run with the phenotype as the dependent variable, and nonmajor allele dosage(s) and a constant\-1 column as predictors. * There is usually an additive effect line for every nonmajor allele, and .TP no such line for the major allele. To omit REF alleles instead of major .TP alleles, add the 'omit\-ref' modifier. (When performing interaction .IP testing, this tends to cause the multicollinearity check to fail for low\-ref\-frequency variants.) .IP * By default, sex (male = 1, female = 2; note that this is a change from .IP PLINK 1.x) is automatically added as a predictor for X chromosome variants, and no others. The 'sex' modifier causes it to be added everywhere (except chrY), while 'no\-x\-sex' excludes it entirely. .IP * The 'log10' modifier causes p\-values to be reported in \fB\-log10\fR(p) form. * 'pheno\-ids' causes the samples used in each set of regressions to be .TP written to an .id file. (When the samples differ on chrX or chrY, .x.id .IP and/or .y.id files are also written.) .IP * The 'genotypic' modifier adds an additive effect/dominance deviation 2df .IP joint test (0\-2 and 0..1..0 coding), while 'hethom' uses 0..0..1 and 0..1..0 coding instead. .IP * 'dominant' and 'recessive' specify a model assuming full dominance or .TP recessiveness, respectively, for the ref allele. I.e. the genotype .IP column is recoded as 0..1..1 or 0..0..1, respectively. .TP * 'interaction' adds genotype x covariate interactions to the model. Note .IP that this tends to produce 'NA' results (due to the multicollinearity check) when the reference allele is 'wrong'; \fB\-\-maj\-ref\fR can be used to enable analysis of those variants. .TP * Additional predictors can be added with \fB\-\-covar\fR. By default, association .IP statistics are reported for all nonconstant predictors; 'hide\-covar' suppresses covariate\-only results, while 'intercept' causes intercepts to be reported. Since running \fB\-\-glm\fR without at least e.g. principal component covariates is usually an analytical mistake, the 'allow\-no\-covars' modifier is now required when you're intentionally running \fB\-\-glm\fR without a covariate file. .IP * By default, if the current phenotype and covariates are such that every .TP regression on a chromosome will fail, PLINK 2 errors out. To just skip .IP the phenotype or chromosome instead, add the 'skip' modifier. .IP * There are now three regression modes for case/control phenotypes: .IP * 'no\-firth' requests PLINK 1.x's behavior, where a NA result is reported .IP when basic logistic regression fails to converge. .IP * 'firth\-fallback' requests logistic regression, followed by Firth .TP regression whenever the logistic regression fails to converge. This is .IP now the default. .IP * 'firth' requests Firth regression all the time. .IP * To add covariates which are not constant across all variants, add the .IP \&'local\-covar=' and 'local\-psam=' modifiers, use full filenames for each, and use either 'local\-pvar=' or 'local\-pos\-cols=' to provide variant ID or position information. Normally, the local\-covar file should have c * n real\-valued columns, where the first c columns correspond to the first sample in the local\-psam file, columns (c+1) to 2c correspond to the second sample, etc.; and the mth line corresponds to the mth nonheader line of the local\-pvar file when there is one. (Variants outside of the local\-pvar file are excluded from the regression.) The local covariates are assigned the names LOCAL1, LOCAL2, etc.; to exclude the last local covariate from the regression (necessary if they are e.g. local ancestry coefficients which sum to 1), add 'local\-omit\-last'. Alternatively, with 'local\-cats=', the local\-covar file is expected to have n columns with integer\-valued entries in [1, k]. (This range is [0, k\-1] with 'local\-cats0='.) These category assignments are expanded into (k\-1) local covariates in the usual manner. When position information is in the local\-covar file, this should be indicated by 'local\-pos\-cols=',,,. \&'local\-haps' indicates that there's one column or column\-group per haplotype instead of per sample; they are averaged by \fB\-\-glm\fR. .IP The main report supports the following column sets: .IP chrom: Chromosome ID. pos: Base\-pair coordinate. (ID is always present, and positioned here.) ref: Reference allele. alt1: Alternate allele 1. alt: All alternate alleles, comma\-separated. (A1 is always present, and positioned here. For multiallelic variants, this column may contain multiple comma\-separated alleles when the result doesn't depend on which allele is A1.) ax: Non\-A1 alleles, comma\-separated. a1count: A1 allele count (can be decimal with dosage data). totallele: Allele observation count (can be higher than \fB\-\-freq\fR value, due .IP to inclusion of het haploids and chrX model). .IP a1countcc: A1 count in cases, then controls (case/control only). totallelecc: Case and control allele observation counts. gcountcc: Genotype hardcall counts (neither\-A1, het\-A1, A1\-A1) in cases, .IP then controls (case/control only). .IP a1freq: A1 allele frequency. a1freqcc: A1 frequency in cases, then controls (case/control only). machr2: Unphased MaCH imputation quality (frequently labeled 'INFO'). firth: Reports whether Firth regression was used (firth\-fallback only). test: Test identifier. (Required unless only one test is run.) nobs: Number of samples in the regression. beta: Regression coefficient (for A1 if additive test). orbeta: Odds ratio for case/control, beta for quantitative traits. .IP (Ignored if 'beta' column set included.) .IP se: Standard error of beta. ci: Bounds of symmetric approximate confidence interval (requires \fB\-\-ci\fR). tz: T\-statistic for linear regression, Wald Z\-score for logistic/Firth. p: Asymptotic p\-value (or \fB\-log10\fR(p)) for T/Z\-statistic. err: Error code for NA results. .IP The default is chrom,pos,ref,alt,firth,test,nobs,orbeta,se,ci,tz,p,err. .HP \fB\-\-score\fR [i] [j] [k] [{header | header\-read}] .IP [{center | variance\-standardize | dominant | recessive}] ['no\-mean\-imputation'] ['se'] ['zs'] ['ignore\-dup\-ids'] [{list\-variants | list\-variants\-zs}] ['cols='] .IP Apply linear scoring system(s) to each sample. The input file should have one line per scored (variant, allele) pair. Variant IDs are read from column #i and allele codes are read from column #j, where i defaults to 1 and j defaults to i+1. * By default, a single column of input coefficients is read from column #k, .TP where k defaults to j+1. (\fB\-\-score\-col\-nums\fR can be used to specify .IP multiple columns.) .IP * 'header\-read' causes the first line of the input file to be treated as a .TP header line containing score names. Otherwise, score(s) are assigned the .IP names 'SCORE1', 'SCORE2', etc.; and 'header' just causes the first line to be entirely ignored. .IP * By default, copies of unnamed alleles contribute zero to score, while .IP missing genotypes contribute an amount proportional to the loaded (via \fB\-\-read\-freq\fR) or imputed allele frequency. To throw out missing observations instead (decreasing the denominator in the final average when this happens), use the 'no\-mean\-imputation' modifier. .IP * You can use the 'center' modifier to shift all genotypes to mean zero, or .IP \&'variance\-standardize' to linearly transform the genotypes to mean\-0, variance\-1. .IP * The 'dominant' modifier causes dosages greater than 1 to be treated as 1, .IP while 'recessive' uses max(dosage \- 1, 0) on diploid chromosomes. ('dominant', 'recessive', and 'variance\-standardize' cannot be used with chrX.) .IP * The 'se' modifier causes the input coefficients to be treated as .IP independent standard errors; in this case, standard errors for the score average/sum are reported. (Note that this will systematically underestimate standard errors when scored variants are in LD.) .IP * By default, \fB\-\-score\fR errors out if a variant ID in the input file appears .TP multiple times in the main dataset. Use the 'ignore\-dup\-ids' modifier to .IP skip them instead (a warning is still printed if such variants are present). .IP * The 'list\-variants[\-zs]' modifier causes variant IDs used for scoring to .IP be written to .sscore.vars[.zst]. .IP The main report supports the following column sets: .IP maybefid: FID, if that column was in the input. fid: Force FID column to be written even when absent in the input. (IID is always present, and positioned here.) maybesid: SID, if that column was in the input. sid: Force SID column to be written even when absent in the input. pheno1: First active phenotype. phenos: All active phenotypes, if any. nallele: Number of nonmissing alleles. denom: Denominator of score average (equal to nallele value when .IP \&'no\-mean\-imputation' specified). .IP dosagesum: Sum of named allele dosages. scoreavgs: Score averages. scoresums: Score sums. .IP The default is maybefid,maybesid,phenos,nallele,dosagesum,scoreavgs. For more sophisticated polygenic risk scoring, we recommend looking at the LDpred (https://github.com/bvilhjal/ldpred ) and PRSice\-2 (https://www.prsice.info/ ) software packages. .HP \fB\-\-variant\-score\fR ['zs'] ['bin' | 'cols='] .IP (alias: \fB\-\-vscore\fR) Apply linear scoring system(s) to each variant. Each reported variant score is the dot product of a sample\-weight vector with the total\-ALT\-dosage vector, with MAF\-based mean imputation applied to missing dosages. Input file format: one line per sample, each starting with an ID and followed by scoring weight(s); it can also have a header line with the sample ID representation and the score name(s). The usual .vscore text report supports the following column sets: .IP chrom: Chromosome ID. pos: Base\-pair coordinate. (ID is always present, and positioned here.) ref: Reference allele. alt1: Alternate allele 1. alt: All alternate alleles, comma\-separated. altfreq: ALT allele frequency used for mean\-imputation. nmiss: Number of missing (and thus mean\-imputed) dosages. nobs: Number of (nonmissing) sample observations. (Variant scores are always present, and positioned here.) .IP Default is chrom,pos,ref,alt. If binary output is requested instead, the main .vscore.bin matrix contains double\-precision floating\-point values, column (score) ID(s) are saved to .vscore.cols, and variant IDs are saved to .vscore.vars[.zst]. .HP \fB\-\-adjust\-file\fR ['zs'] ['gc'] ['cols='] .IP ['log10'] ['input\-log10'] ['test='] .IP Given a file with unfiltered association test results, report some basic multiple\-testing corrections, sorted in increasing\-p\-value order. * 'gc' causes genomic\-controlled p\-values to be used in the formulas. .TP (This tends to be overly conservative. We note that LD Score regression .IP usually does a better job of calibrating lambda; see Lee JJ, McGue M, Iacono WG, Chow CC (2018) The accuracy of LD Score regression as an estimator of confounding and genetic correlations in genome\-wide association studies.) .IP * 'log10' causes negative base 10 logs of p\-values to be reported, instead .TP of raw p\-values. \&'input\-log10' specifies that the input file contains .HP \fB\-log10\fR(p) values. .IP * If the input file contains multiple tests per variant which are .IP distinguished by a 'TEST' column (true for \fB\-\-linear\fR/\-\-logistic/\-\-glm), you must use 'test=' to select the test to process. .IP The following column sets are supported: .IP chrom: Chromosome ID. pos: Base\-pair coordinate. (ID is always present, and positioned here.) ref: Reference allele. alt1: Alternate allele 1. alt: All alternate alleles, comma\-separated. a1: Tested allele. (Omitted if missing from input file.) unadj: Unadjusted p\-value. gc: Devlin & Roeder (1999) genomic control corrected p\-value (additive .IP models only). .IP qq: P\-value quantile. bonf: Bonferroni correction. holm: Holm\-Bonferroni (1979) adjusted p\-value. sidakss: Sidak single\-step adjusted p\-value. sidaksd: Sidak step\-down adjusted p\-value. fdrbh: Benjamini & Hochberg (1995) step\-up false discovery control. fdrby: Benjamini & Yekutieli (2001) step\-up false discovery control. .IP Default set is chrom,a1,unadj,gc,bonf,holm,sidakss,sidaksd,fdrbh,fdrby. .HP \fB\-\-genotyping\-rate\fR ['dosage'] .IP Report genotyping rate in log (this was automatic in PLINK 1.x). .HP \fB\-\-pgen\-info\fR .IP Reports basic information about a .pgen file. .HP \fB\-\-validate\fR .IP Validates all variant records in a .pgen file. .HP \fB\-\-zst\-decompress\fR <.zst file> [output filename] .IP (alias: \fB\-\-zd\fR) Decompress a Zstd\-compressed file. If no output filename is specified, the file is decompressed to standard output. This cannot be used with any other flags, and does not cause a log file to be generated. .PP The following other flags are supported. .TP \fB\-\-script\fR : Include command\-line options from file. .TP \fB\-\-rerun\fR [log] : Rerun commands in log (default 'plink2.log'). .TP \fB\-\-version\fR : Display only version number before exiting. .TP \fB\-\-silent\fR : Suppress regular output to console. (Error\-output is not suppressed.) .TP \fB\-\-double\-id\fR : Set both FIDs and IIDs to the VCF/.bgen sample ID. .TP \fB\-\-const\-fid\fR [ID] : Set all FIDs to the given constant. If '0' (the default), no FID column is created. .TP \fB\-\-id\-delim\fR [d] : Normally parses single\-delimiter sample IDs as , and double\-delimiter IDs as ; default delimiter is '_'. \fB\-\-id\-delim\fR can no longer be used with \fB\-\-double\-id\fR/\-\-const\-fid; it will error out if any ID lacks the delimiter. .TP \fB\-\-idspace\-to\fR : Convert spaces in VCF/.bgen sample IDs to the given character. .TP \fB\-\-iid\-sid\fR : Make \fB\-\-id\-delim\fR and \fB\-\-sample\-diff\fR interpret two\-token sample IDs as IID\-SID instead of FID\-IID. .TP \fB\-\-vcf\-require\-gt\fR : Skip variants with no GT field. .TP \fB\-\-vcf\-min\-gq\fR : No\-call genotypes when GQ is present and below the threshold. .TP \fB\-\-vcf\-max\-dp\fR : No\-call genotypes when DP is present and above/below .TP \fB\-\-vcf\-min\-dp\fR the threshold. .TP \fB\-\-vcf\-half\-call\fR : Specify how '0/.' and similar VCF GT values should be handled. The following four modes are supported: * 'error'/'e' (default) errors out and reports line #. * 'haploid'/'h' treats them as haploid calls. * 'missing'/'m' treats them as missing. * 'reference'/'r' treats the missing value as 0. .TP \fB\-\-oxford\-single\-chr\fR : Specify single\-chromosome .gen/.bgen file with no useful chromosome info inside. .TP \fB\-\-missing\-code\fR [string list] : Comma\-delimited list of missing phenotype .TP (alias: \fB\-\-missing_code\fR) values for Oxford\-format import (default \&'NA'). .TP \fB\-\-hard\-call\-threshold\fR : When importing dosage data, a hardcall is normally saved when the distance from the nearest hardcall, defined as .TP 0.5 * sum_i |x_i \- round(x_i)| (where the x_i's are 0..2 allele dosages), is not greater than 0.1. You can adjust this threshold by providing a numeric argument to \fB\-\-hard\-call\-threshold\fR. You can also use this with \fB\-\-make\-[b]pgen\fR to alter the saved hardcalls while leaving the dosages untouched, or \fB\-\-make\-bed\fR to tweak hardcall export. .TP \fB\-\-dosage\-erase\-threshold\fR : \fB\-\-hard\-call\-threshold\fR normally preserves the original dosages, and several PLINK 2 commands use them when they're available. Use \fB\-\-dosage\-erase\-threshold\fR to make PLINK 2 erase dosages and keep only hardcalls when distance\-from\-hardcall <= the given level. .TP \fB\-\-import\-dosage\-certainty\fR : The PLINK 2 file format currently supports a single dosage for each allele. Some other dosage file formats include a separate probability for every possible genotype, e.g. {P(0/0)=0.2, P(0/1)=0.52, P(1/1)=0.28}, a highly uncertain call that is nevertheless treated as a hardcall under \&'\-\-hard\-call\-threshold 0.1'. To make PLINK 2 treat a dosage as missing whenever the largest probability is less than a threshold, use \fB\-\-import\-dosage\-certainty\fR. .TP \fB\-\-input\-missing\-genotype\fR : '.' is always interpreted as a missing genotype code in input files. By default, '0' also is; you can change this second missing code with \fB\-\-input\-missing\-genotype\fR. .TP \fB\-\-allow\-extra\-chr\fR : Permit unrecognized chromosome codes (alias \fB\-\-aec\fR). .HP \fB\-\-chr\-set\fR ['no\-x'] ['no\-y'] ['no\-xy'] ['no\-mt'] : .TP Specify a nonhuman chromosome set. The first parameter sets the number of .IP diploid autosome pairs if positive, or haploid chromosomes if negative. Given diploid autosomes, the remaining modifiers indicate the absence of the named non\-autosomal chromosomes. .HP \fB\-\-cow\fR/\-\-dog/\-\-horse/\-\-mouse/\-\-rice/\-\-sheep : Shortcuts for those species. .TP \fB\-\-autosome\-num\fR : Alias for '\-\-chr\-set no\-y no\-xy no\-mt'. .TP \fB\-\-human\fR : Explicitly specify human chromosome set, and make output .pvar/VCF files include a ##chrSet header line. (.pvar/VCF output files automatically include ##chrSet when a nonhuman set is specified.) .TP \fB\-\-chr\-override\fR ['file'] : By default, if \fB\-\-chr\-set\fR/\-\-autosome\-num/\-\-cow/etc. conflicts with an input file ##chrSet header line, PLINK 2 will error out. \fB\-\-chr\-override\fR with no argument causes the command line to take precedence; '\-\-chr\-override file' defers to the file. .TP \fB\-\-var\-min\-qual\fR : Skip variants with low/missing QUAL. .TP \fB\-\-var\-filter\fR [exception(s)...] : Skip variants which have FILTER failures. .HP \fB\-\-extract\-if\-info\fR : Exclude variants which don't/do satisfy .TP \fB\-\-exclude\-if\-info\fR a comparison predicate on an INFO key, .TP (aliases: \fB\-\-extract\-if\fR, e.g. .TP \fB\-\-exclude\-if\fR) \fB\-\-extract\-if\-info\fR "VT == SNP" .TP Unless the operator is !=, the predicate always evaluates to false when the key is missing. .TP \fB\-\-require\-info\fR : Exclude variants based on nonexistence .TP \fB\-\-require\-no\-info\fR or existence of an INFO key. "=." is treated as nonexistence. .HP \fB\-\-extract\-col\-cond\fR [valcol] [IDcol] [skip] : .HP \fB\-\-extract\-col\-cond\-match\fR <(sub)string(s)...> .HP \fB\-\-extract\-col\-cond\-mismatch\fR <(sub)string(s)...> .HP \fB\-\-extract\-col\-cond\-substr\fR .HP \fB\-\-extract\-col\-cond\-min\fR .HP \fB\-\-extract\-col\-cond\-max\fR : .IP Exclude all variants without a value\-column entry satisfying a condition. * By default, values are read from column 2 of the file, and variant IDs .IP are read from column 1. .IP * Three types of conditions are supported: .IP * When \fB\-\-extract\-col\-cond\-match\fR is specified without .HP \fB\-\-extract\-col\-cond\-substr\fR, the value is checked for equality with the .TP given strings, and kept iff one of them matches. Similarly, .HP \fB\-\-extract\-col\-cond\-mismatch\fR without \fB\-\-extract\-col\-cond\-substr\fR causes .IP the variant to be kept iff the value matches none of the given strings. .IP * When \fB\-\-extract\-col\-cond\-match\fR and/or \fB\-mismatch\fR are specified with .HP \fB\-\-extract\-col\-cond\-substr\fR, the variant is kept iff none of the .HP \fB\-\-extract\-col\-cond\-mismatch\fR substrings are contained in the value, and .IP either \fB\-\-extract\-col\-cond\-match\fR was unspecified or at least one of its substrings is contained. .IP * Otherwise, the value is interpreted as a number, and the variant is .IP kept if the number is in [, ] (default min=0, max=DBL_MAX). .TP \fB\-\-pheno\fR ['iid\-only'] : Specify additional phenotype/covariate file. Comma\-delimited files with a header line are now permitted. .TP \fB\-\-pheno\-name\fR : Only load the designated phenotype(s) from the \fB\-\-pheno\fR (if one was specified) or .psam (if no \fB\-\-pheno\fR) file. Separate multiple names with spaces or commas, and use dashes to designate ranges. .TP \fB\-\-pheno\-col\-nums\fR <#...> : Only load the phenotype(s) in the designated column number(s) from the \fB\-\-pheno\fR file. .TP \fB\-\-no\-psam\-pheno\fR : Ignore phenotype(s) in .psam/.fam file. .TP \fB\-\-strict\-sid0\fR : By default, if there is no SID column in the .psam/.fam (or \fB\-\-update\-ids\fR) file, but there is one in another input file (for e.g. \fB\-\-keep\fR/\-\-remove), the latter SID column is ignored; sample IDs are considered matching as long as FID and IID are equal (with missing FID treated as '0'). If you also want to require SID = '0' for a sample ID match in this situation, add \fB\-\-strict\-sid0\fR. .TP \fB\-\-input\-missing\-phenotype\fR : Set nonzero number to treat as a missing pheno/covar in input files (default \fB\-9\fR). .TP \fB\-\-no\-input\-missing\-phenotype\fR : Don't treat any nonzero number as a missing pheno/covar. ('NA'/'nan' are still treated as missing.) .TP \fB\-\-1\fR : Expect case/control phenotypes in input files to be coded as 0 = control, 1 = case, instead of the usual 0 = missing, 1 = ctrl, 2 = case. (Unlike PLINK 1.x, this does not force all phenotypes to be interpreted as case/ctrl.) .TP \fB\-\-missing\-catname\fR : Set missing\-categorical\-phenotype string (case\-sensitive, default 'NONE'). .TP \fB\-\-covar\fR ['iid\-only'] : Specify additional covariate file. Comma\-delimited files with a header line are now permitted. .TP \fB\-\-covar\-name\fR : Only load the designated covariate(s) from the \fB\-\-covar\fR (if one was specified), \fB\-\-pheno\fR (if no \fB\-\-covar\fR), or .psam (if no \fB\-\-covar\fR or \fB\-\-pheno\fR) file. .TP \fB\-\-covar\-col\-nums\fR <#...> : Only load the covariate(s) in the designated column number(s) from the \fB\-\-covar\fR (if one was specified) or \fB\-\-pheno\fR (if no \fB\-\-covar\fR) file. .TP \fB\-\-within\fR [new pheno name] : Import a PLINK 1.x categorical phenotype. (Phenotype name defaults to 'CATPHENO'.) * If any numeric values are present, ALL .TP values must be numeric. In that case, 'C' .TP is added in front of all category names. * 'NA' is treated as a missing value. .TP \fB\-\-mwithin\fR : Load \fB\-\-within\fR categories from column n+2. .TP \fB\-\-family\fR [new pheno name] : Create a categorical phenotype from FID. Restrictions on and handling of numeric values are the same as for \fB\-\-within\fR. .TP \fB\-\-family\-missing\-catname\fR : Make \fB\-\-family\fR treat the specified FID as missing. .TP \fB\-\-keep\fR : Exclude all samples not named in a file. .TP \fB\-\-remove\fR : Exclude all samples named in a file. .TP \fB\-\-keep\-fam\fR : Exclude all families not named in a file. .TP \fB\-\-remove\-fam\fR : Exclude all families named in a file. .HP \fB\-\-extract\fR [{bed0 | bed1}] : Usually excludes all variants (not) named .TP \fB\-\-exclude\fR [{bed0 | bed1}] in the given file(s). When multiple files are named, they are concatenated. With the 'bed0' or 'bed1' modifier, variants outside/inside the positional ranges in the interval\-BED file(s) are excluded instead. 'bed0' tells PLINK 2 to assume the interval bounds follow the UCSC 0\-based half\-open convention, while 'bed1' (equivalent to PLINK 1.9 'range') specifies 1\-based fully\-closed. .TP \fB\-\-extract\-intersect\fR [{bed0 | bed1}] : Just like \fB\-\-extract\fR, except that a variant must be in the intersection, rather than just the union, of the files to remain. .TP \fB\-\-bed\-border\-bp\fR : Stretch BED intervals by the given amount on each .TP \fB\-\-bed\-border\-kb\fR side. .TP \fB\-\-keep\-cats\fR : These can be used individually or in combination .TP \fB\-\-keep\-cat\-names\fR to define a list of categories to keep; all samples not in one of the named categories are excluded. Use spaces to separate category names for \fB\-\-keep\-cat\-names\fR. Use the \fB\-\-missing\-catname\fR value (default 'NONE') to refer to the group of uncategorized samples. .TP \fB\-\-keep\-cat\-pheno\fR : If more than one categorical phenotype is loaded, or you wish to filter on a categorical covariate, \fB\-\-keep\-cat\-pheno\fR must be used to specify which phenotype/covariate \fB\-\-keep\-cats\fR and \fB\-\-keep\-cat\-names\fR apply to. .HP \fB\-\-remove\-cats\fR : Exclude all categories named in the file. .HP \fB\-\-remove\-cat\-names\fR <...> : Exclude named categories. .HP \fB\-\-remove\-cat\-pheno\fR : Specify pheno for \fB\-\-remove\-cats\fR/remove\-cat\-names. .TP \fB\-\-split\-cat\-pheno\fR [{omit\-most | omit\-last}] ['covar\-01'] [cat. pheno/covar name(s)...] : .IP Split n\-category phenotype(s) into n (or n\-1, with 'omit\-most'/'omit\-last') binary phenotypes, with names of the form =. (As a consequence, affected phenotypes and categories are not permitted to contain the '=' character.) * This happens after all sample filters. * If no phenotype or covariate names are provided, all categorical .IP phenotypes (but not covariates) are processed. .TP * By default, generated covariates are coded as 1=false, 2=true. To code .IP them as 0=false, 1=true instead, add the 'covar\-01' modifier. .TP \fB\-\-loop\-cats\fR : Run variant filters and subsequent operations on just the samples in the first category; then just the samples in the second category; and so on, for all categories in the named categorical phenotype. .TP \fB\-\-no\-id\-header\fR ['iid\-only'] : Don't include a header line in .id output files. This normally forces two\-column FID/IID output; add 'iid\-only' to force just single\-column IID. .HP \fB\-\-variance\-standardize\fR [pheno/covar name(s)...] .HP \fB\-\-covar\-variance\-standardize\fR [covar name(s)...] : .IP Linearly transform named covariates (and quantitative phenotypes, if \fB\-\-variance\-standardize\fR) to mean\-zero, variance 1. If no arguments are provided, all possible phenotypes/covariates are affected. This is frequently necessary to prevent multicollinearity when dealing with covariates where abs(mean) is much larger than abs(standard deviation), such as year of birth. .TP \fB\-\-quantile\-normalize\fR [...] : Force named covariates and quantitative .TP \fB\-\-pheno\-quantile\-normalize\fR [...] phenotypes to a N(0,1) distribution, .TP \fB\-\-covar\-quantile\-normalize\fR [...] preserving only the original rank orders. .TP \fB\-\-chr\fR : Exclude all variants not on the given chromosome(s). Valid choices for humans are 0 (unplaced), 1\-22, X, Y, XY, MT, PAR1, and PAR2. Separate multiple chromosomes with spaces and/or commas, and use a dash (no adjacent spaces permitted) to denote a range, e.g. \&'\-\-chr 1\-4, 22, par1, x, par2'. .TP \fB\-\-not\-chr\fR <...> : Reverse of \fB\-\-chr\fR (exclude variants on listed chromosomes). .TP \fB\-\-autosome\fR : Exclude all non\-autosomal variants. .TP \fB\-\-autosome\-par\fR : Exclude all non\-autosomal variants, except those in a pseudo\-autosomal region. .TP \fB\-\-snps\-only\fR ['just\-acgt'] : Exclude non\-SNP variants. By default, SNP = all allele codes are single\-character (so multiallelic variants with a mix of SNPs and non\-SNPs are excluded; split your variants first if that's a problem). The 'just\-acgt' modifier restricts SNP codes to {A,C,G,T,a,c,g,t,}. .TP \fB\-\-from\fR : Use ID(s) to specify a variant range to load. When used .TP \fB\-\-to\fR together, both variants must be on the same chromosome. (\fB\-\-snps\fR can be used to specify intervals which cross chromosome boundaries.) .TP \fB\-\-snp\fR : Specify a single variant to load. .HP \fB\-\-exclude\-snp\fR : Specify a single variant to exclude. .TP \fB\-\-window\fR : With \fB\-\-snp\fR/\-\-exclude\-snp, loads/excludes all variants within half the specified kb distance of the named one. .TP \fB\-\-from\-bp\fR : Use base\-pair coordinates to define a variant range to .TP \fB\-\-to\-bp\fR load. .TP \fB\-\-from\-kb\fR * You must use these with \fB\-\-chr\fR, specifying a single .TP \fB\-\-to\-kb\fR chromosome. .TP \fB\-\-from\-mb\fR * Decimals and negative numbers are permitted. .TP \fB\-\-to\-mb\fR * The \fB\-\-to\-bp\fR(/\-kb/\-mb) position is no longer permitted to be smaller than the \fB\-\-from\-bp\fR position. .TP \fB\-\-snps\fR : Use IDs to specify variant range(s) to load or .TP \fB\-\-exclude\-snps\fR <...> exclude. E.g. '\-\-snps rs1111\-rs2222, rs3333, rs4444'. .TP \fB\-\-force\-intersect\fR : PLINK 2 normally errors out when multiple variant inclusion filters (\fB\-\-extract\fR, \fB\-\-extract\-col\-cond\fR, \fB\-\-extract\-intersect\fR, \fB\-\-from\fR/\-\-to, \fB\-\-from\-bp\fR/\-\-to\-bp, \fB\-\-snp\fR, \fB\-\-snps\fR) are specified. \fB\-\-force\-intersect\fR allows the run to proceed; the set intersection will be taken. .TP \fB\-\-thin\fR

: Randomly remove variants, retaining each with prob. p. .TP \fB\-\-thin\-count\fR : Randomly remove variants until n of them remain. .TP \fB\-\-bp\-space\fR : Remove variants so that each pair is no closer than the given bp distance. .TP \fB\-\-thin\-indiv\fR

: Randomly remove samples, retaining with prob. p. .HP \fB\-\-thin\-indiv\-count\fR : Randomly remove samples until n of them remain. .TP \fB\-\-keep\-col\-match\fR : Exclude all samples without a 3rd column entry in the given file exactly matching one of the given strings. (Separate multiple strings with spaces.) .HP \fB\-\-keep\-col\-match\-name\fR : Check column with given name instead. .TP \fB\-\-keep\-col\-match\-num\fR : Check nth column instead. .HP \fB\-\-geno\fR [val] [{dosage | hh\-missing}] .HP \fB\-\-mind\fR [val] [{dosage | hh\-missing}] : .IP Exclude variants (\fB\-\-geno\fR) and/or samples (\fB\-\-mind\fR) with missing call frequencies greater than a threshold (default 0.1). (Note that the default threshold is only applied if \fB\-\-geno\fR/\-\-mind is invoked without an argument; when \fB\-\-geno\fR/\-\-mind is not invoked, no missing call frequency ceiling is enforced at all. Other inclusion/exclusion default thresholds work the same way.) By default, when a dosage is present but a hardcall is not, the genotype is treated as missing; add the 'dosage' modifier to treat this case as nonmissing. Alternatively, you can use 'hh\-missing' to also treat heterozygous haploid calls as missing. .HP \fB\-\-require\-pheno\fR [name(s)...] : Remove samples missing any of the named .TP \fB\-\-require\-covar\fR [name(s)...] phenotype(s)/covariate(s). If no arguments are provided, all phenotype(s)/covariate(s) must be present. .TP \fB\-\-maf\fR [freq] [mode] : Exclude variants with allele frequency lower than a .TP (alias: \fB\-\-min\-af\fR) threshold (default 0.01). By default, the nonmajor allele frequency is used; the other supported modes are 'nref' (non\-reference), 'alt1', and 'minor' (least frequent). bcftools freq:mode notation is permitted. .HP \fB\-\-max\-maf\fR [mode] : Exclude variants with MAF greater than the .TP (alias: \fB\-\-max\-af\fR) threshold. .TP \fB\-\-mac\fR [mode] : Exclude variants with allele dosage lower than the .TP (alias: \fB\-\-min\-ac\fR) given threshold. .TP \fB\-\-max\-mac\fR [mode] : Exclude variants with allele dosage greater than .TP (alias: \fB\-\-max\-ac\fR) the given threshold. .TP \fB\-\-maf\-succ\fR : Rule of succession allele frequency estimation (used in EIGENSOFT). Given j observations of one allele and k observations of the other for a biallelic variant, infer allele frequencies of (j+1) / (j+k+2) and (k+1) / (j+k+2), rather than the default j / (j+k) and k / (j+k). Note that this does not affect \fB\-\-freq\fR's output. .TP \fB\-\-min\-alleles\fR : Exclude variants with fewer than the given # of alleles. (When a variant has exactly one ALT allele, and it's a missing\-code, it's excluded by "\-\-min\-alleles 2".) .HP \fB\-\-max\-alleles\fR : Exclude variants with more than the given # of alleles. .TP \fB\-\-read\-freq\fR : Load allele frequency estimates from the given \fB\-\-freq\fR or \fB\-\-geno\-counts\fR (or PLINK 1.9 \fB\-\-freqx\fR) report, instead of imputing them from the immediate dataset. .HP \fB\-\-hwe\fR

['midp'] ['keep\-fewhet'] : .IP Exclude variants with Hardy\-Weinberg equilibrium exact test p\-values below a threshold. * By default, only founders are considered. * chrX p\-values are now computed using Graffelman and Weir's method. * For variants with k alleles with k>2, k separate 'biallelic' tests are .IP performed, and the variant is filtered out if any of them fail. .IP * With 'keep\-fewhet', variants which fail the test in the too\-few\-hets .TP direction are not excluded. On chrX, this uses the ratio between the .IP Graffelman/Weir p\-value and the female\-only p\-value. .IP * There is currently no special handling of case/control phenotypes. .TP \fB\-\-mach\-r2\-filter\fR [min] [max] : Exclude variants with MaCH imputation quality metric less than min or greater than max (defaults 0.1 and 2.0). (Monomorphic variants, with r2 = nan, are not excluded.) * This is NOT identical to the R2 metric .TP reported by Minimac3 0.1.13+; see below. * If a single argument is provided, it is .TP treated as the minimum. * The metric is not computed on chrX and MT. .TP \fB\-\-minimac3\-r2\-filter\fR [max] : Compute Minimac3 R2 values from scratch, and exclude variants with R2 less than min or (if max is provided) greater than max. * Note that this requires phased\-dosage .TP data for all samples and variants; otherwise this will systematically underestimate imputation quality, since unphased hardcalls/dosages are treated as if they were maximally uncertain. (Use \fB\-\-extract\-if\-info\fR/\-\-exclude\-if\-info to filter on precomputed Minimac3 R2 in a VCF/.pvar INFO column.) .TP \fB\-\-keep\-females\fR : Exclude male and unknown\-sex samples. .TP \fB\-\-keep\-males\fR : Exclude female and unknown\-sex samples. .TP \fB\-\-keep\-nosex\fR : Exclude all known\-sex samples. .TP \fB\-\-remove\-females\fR : Exclude female samples. .TP \fB\-\-remove\-males\fR : Exclude male samples. .TP \fB\-\-remove\-nosex\fR : Exclude unknown\-sex samples. .TP \fB\-\-keep\-founders\fR : Exclude nonfounder samples. .HP \fB\-\-keep\-nonfounders\fR : Exclude founder samples. .HP \fB\-\-keep\-if\fR : Exclude samples which don't/do satisfy a .TP \fB\-\-remove\-if\fR comparison predicate, e.g. \fB\-\-keep\-if\fR "PHENO1 == case" .TP Unless the operator is !=, the predicate always evaluates to false when the phenotype/covariate is missing. .TP \fB\-\-nonfounders\fR : Include nonfounders in allele freq/HWE calculations. .TP \fB\-\-bad\-freqs\fR : When PLINK 2 needs decent allele frequencies, it normally errors out if they aren't provided by \fB\-\-read\-freq\fR and less than 50 founders are available to impute them from. Use \fB\-\-bad\-freqs\fR to force PLINK 2 to proceed in this case. .TP \fB\-\-bad\-ld\fR : PLINK 2 normally errors out when it needs to estimate LD between variants, but there are less than 50 founders to estimate from. Use \fB\-\-bad\-ld\fR to force PLINK 2 to proceed. .TP \fB\-\-export\-allele\fR : With \fB\-\-export\fR A/A\-transpose/AD, count alleles named in the file, instead of REF alleles. .TP \fB\-\-output\-chr\fR : Set chromosome coding scheme in output files by providing the desired human mitochondrial code. Options are '26', 'M', 'MT', '0M', 'chr26', 'chrM', and 'chrMT'; default is now 'MT' (note that this is a change from PLINK 1.x, which defaulted to '26'). .TP \fB\-\-output\-missing\-genotype\fR : Set the code used to represent missing genotypes in output files (default '.'). .TP \fB\-\-output\-missing\-phenotype\fR : Set the string used to represent missing phenotypes in output files (default 'NA'). .TP \fB\-\-sort\-vars\fR [mode] : Sort variants by chromosome, then position, then ID. The following string orders are supported: * 'natural'/'n': Natural sort (default). * 'ascii'/'a': ASCII. This must be used with \fB\-\-make\-[b]pgen\fR/\-\-make\-bed. .TP \fB\-\-set\-hh\-missing\fR ['keep\-dosage'] : Make \fB\-\-make\-[b]pgen\fR/\-\-make\-bed set non\-MT heterozygous haploid hardcalls, and all female chrY calls, to missing. (Unlike PLINK 1.x, this treats unknown\-sex chrY genotypes like males, not females.) By default, all associated dosages are also erased; use 'keep\-dosage' to keep them all. .TP \fB\-\-set\-mixed\-mt\-missing\fR ['keep\-dosage'] : Make \fB\-\-make\-[b]pgen\fR/\-\-make\-bed set mixed MT hardcalls to missing. .HP \fB\-\-split\-par\fR : Changes chromosome code of all X chromosome .TP \fB\-\-split\-par\fR variants with bp position <= bp1 to PAR1, and those with position >= bp2 to PAR2. The following build codes are supported as shorthand: * 'b36'/'hg18' = NCBI 36, 2709521/154584237 * 'b37'/'hg19' = GRCh37, 2699520/154931044 * 'b38'/'hg38' = GRCh38, 2781479/155701383 .TP \fB\-\-merge\-par\fR : Merge PAR1/PAR2 back with X. Requires PAR1 to be positioned immediately before X, and PAR2 to be immediately after X. (Should *not* be used with "\-\-export vcf", since it causes male homozygous/missing calls in PAR1/PAR2 to be reported as haploid.) .TP \fB\-\-merge\-x\fR : Merge XY back with X. This usually has to be combined with \fB\-\-sort\-vars\fR. .TP \fB\-\-set\-missing\-var\-ids\fR : Given a template string with a '@' where the .TP \fB\-\-set\-all\-var\-ids\fR chromosome code should go and '#' where the bp coordinate belongs, \fB\-\-set\-missing\-var\-ids\fR assigns chromosome\-and\-bp\-based IDs to unnamed variants, while \fB\-\-set\-all\-var\-ids\fR resets all IDs. You may also use '$r'/'$a' to refer to the ref and alt1 alleles, or '$1'/'$2' to refer to them in alphabetical order. .TP \fB\-\-var\-id\-multi\fR : Specify alternative templates for multiallelic .TP \fB\-\-var\-id\-multi\-nonsnp\fR variants. ('$a' and '$1'/'$2' should be avoided here, though they're technically still allowed.) .HP \fB\-\-new\-id\-max\-allele\-len\fR [{error | missing | truncate}] : .IP Specify maximum number of leading characters from allele codes to include in new variant IDs, and behavior on longer codes (defaults 23, error). .TP \fB\-\-missing\-var\-code\fR : Change unnamed variant code for \fB\-\-rm\-dup\fR, \fB\-\-set\-\fR{missing|all}\-var\-ids, and \fB\-\-recover\-var\-ids\fR (default '.'). .TP \fB\-\-update\-map\fR [bpcol] [IDcol] [skip] : Update variant bp positions. .HP \fB\-\-update\-name\fR [newcol] [oldcol] [skip] : Update variant IDs. .HP \fB\-\-recover\-var\-ids\fR ['strict\-bim\-order'] [{rigid | force}] ['partial'] : .TP Undo \fB\-\-set\-all\-var\-ids\fR, given the original .pvar/VCF/.bim file. Original .IP IDs are looked up by position and allele codes. * By default, if the original\-ID file is a .bim, allele order is ignored. .IP Use 'strict\-bim\-order' to force A1=ALT, A2=REF. .IP * If any variant has multiple matching records in the original\-ID file, and .IP the IDs conflict, \fB\-\-recover\-var\-ids\fR writes the affected (current) ID(s) to .recoverid.dup, and normally errors out. If the original\-ID file has the same number of variants in the same order, you can still recover the old IDs with the 'rigid' modifier in this case. Alternatively, to proceed and assign the missing\-ID code to these variants, add the 'force' modifier. (The .recoverid.dup file is still written when 'rigid' or 'force' is specified.) .IP * \fB\-\-recover\-var\-ids\fR normally expects to replace all variant IDs, and errors .TP out if any are left untouched. Add the 'partial' modifier when you .IP actually want to update just a proper subset. .TP \fB\-\-update\-alleles\fR : Update variant allele codes. .TP \fB\-\-update\-ids\fR : Update sample IDs. .TP \fB\-\-update\-parents\fR : Update parental IDs. .HP \fB\-\-update\-sex\fR ['col\-num='] ['male0'] : .IP Update sex information. * By default, if there is a header line starting with '#FID'/'#IID', sex is .IP loaded from the first column titled 'SEX' (any capitalization); otherwise, column 3 is assumed. Use 'col\-num=' to force a column number. .TP * Only the first character in the sex column is processed. By default, .IP \&'1'/'M'/'m' is interpreted as male, '2'/'F'/'f' is interpreted as female, and '0'/'N' is interpreted as unknown\-sex. To change this to '0'/'M'/'m' = male, '1'/'F'/'f' = female, anything else other than '2' = unknown\-sex, add 'male0'. .TP \fB\-\-real\-ref\-alleles\fR : Treat A2 alleles in a PLINK 1.x fileset as actual REF alleles; otherwise they're marked as provisional. .TP \fB\-\-maj\-ref\fR ['force'] : Set major alleles to reference, like PLINK 1.x automatically did. (Note that this is now opt\-in rather than opt\-out; \fB\-\-keep\-allele\-order\fR is no longer necessary to prevent allele\-swapping.) * This can only be used in runs with .TP \fB\-\-make\-bed\fR/\-\-make\-[b]pgen/\-\-export and no other commands. .TP * By default, this only affects variants marked as having 'provisional' reference alleles. Add 'force' to apply this to all variants. .IP * All new reference alleles are marked as provisional. .HP \fB\-\-ref\-allele\fR ['force'] [refcol] [IDcol] [skip] .HP \fB\-\-alt1\-allele\fR ['force'] [alt1col] [IDcol] [skip] : .IP These set the alleles specified in the file to ref (\fB\-\-ref\-allele\fR) or alt1 (\fB\-\-alt1\-allele\fR). They can be combined in the same run. * These can only be used in runs with \fB\-\-make\-bed\fR/\-\-make\-[b]pgen/\-\-export .IP and no other commands. .IP * "\-\-ref\-allele 4 3 '#'", which scrapes reference allele .IP assignments from a VCF file, is especially useful. .IP * By default, these error out when asked to change a 'known' reference .TP allele. Add 'force' to permit that (when e.g. switching to a new .IP reference genome). .IP * When \fB\-\-alt1\-allele\fR changes the previous ref allele to alt1, the previous .IP alt1 allele is set to reference and marked as provisional. .TP \fB\-\-ref\-from\-fa\fR ['force'] : This sets reference alleles from the \fB\-\-fa\fR file when it can be done unambiguously (note that it's never possible for deletions or some insertions). By default, it errors out when asked to change a \&'known' reference allele; add the 'force' modifier to permit that. .TP \fB\-\-normalize\fR ['list'] : Left\-normalize all variants, using the \fB\-\-fa\fR file. .TP (alias: \fB\-\-norm\fR) (Assumes no differences in capitalization.) The \&'list' modifier causes a list of affected variant IDs to be written to .normalized. .TP \fB\-\-indiv\-sort\fR [f] : Specify sample ID sort order for merge and \fB\-\-make\-[b]pgen\fR/\-\-make\-bed. The following four modes are supported: * 'none'/'0' keeps samples in the order they were .TP loaded. Default for non\-merge. .TP * 'natural'/'n' invokes "natural sort", e.g. \&'id2' < 'ID3' < 'id10'. Default when merging. .TP * 'ascii'/'a' sorts in ASCII order, e.g. \&'ID3' < 'id10' < 'id2'. .TP * 'file'/'f' uses the order in the given file (named in the last argument). .TP \fB\-\-king\-table\-filter\fR : Specify minimum kinship coefficient for inclusion in \fB\-\-make\-king\-table\fR report. .TP \fB\-\-king\-table\-subset\fR [kmin] : Restrict current \fB\-\-make\-king\-table\fR run to sample pairs listed in the given .kin0 file. If a second argument is provided, only sample pairs with kinship >= that threshold (in the input .kin0) are processed. .HP \fB\-\-condition\fR [{dominant | recessive}] ['multiallelic'] .HP \fB\-\-condition\-list\fR [{dominant | recessive}] ['multiallelic'] : .IP Add the given variant, or all variants in the given file, as \fB\-\-glm\fR covariates. By default, this errors out if any of the variants are multiallelic; add the 'multiallelic' ('m' for short) modifier to allow them. They'll effectively be split against the major allele (unless \fB\-\-glm\fR's 'omit\-ref' modifier was specified), and all induced covariate names\-\-even for biallelic variants\-\-will have an underscore followed by the allele code at the end. .TP \fB\-\-parameters\fR <...> : Include only the given covariates/interactions in the \fB\-\-glm\fR model, identified by a list of 1\-based indices and/or ranges of them. .TP \fB\-\-tests\fR <...> : Perform a (joint) test on the specified term(s) in the .TP \fB\-\-tests\fR all \fB\-\-glm\fR model, identified by 1\-based indices and/or ranges of them. * Note that, when \fB\-\-parameters\fR is also present, the .TP indices refer to the terms remaining AFTER pruning by \fB\-\-parameters\fR. .IP * You can use '\-\-tests all' to include all terms. .TP \fB\-\-vif\fR : Set VIF threshold for \fB\-\-glm\fR multicollinearity check (default 50). (This is no longer skipped for case/control phenotypes.) .TP \fB\-\-max\-corr\fR : Skip \fB\-\-glm\fR regression when the absolute value of the correlation between two predictors exceeds this value (default 0.999). .TP \fB\-\-xchr\-model\fR : Set the chrX \fB\-\-glm\fR/\-\-condition[\-list]/\-\-[v]score model. * '0' = skip chrX. * '1' = add sex as a covar on chrX, code males 0..1. * '2' (default) = chrX sex covar, code males 0..2. (Use the \fB\-\-glm\fR 'interaction' modifier to test for interaction between genotype and sex.) .HP \fB\-\-adjust\fR ['zs'] ['gc'] ['log10'] ['cols='] : .IP For each association test in this run, report some basic multiple\-testing corrections, sorted in increasing\-p\-value order. Modifiers work the same way as they do on \fB\-\-adjust\-file\fR. .TP \fB\-\-lambda\fR : Set genomic control lambda for \fB\-\-adjust[\-file]\fR. .TP \fB\-\-adjust\-chr\-field\fR : Set \fB\-\-adjust\-file\fR input field names. When .TP \fB\-\-adjust\-pos\-field\fR multiple arguments are given to these flags, .TP \fB\-\-adjust\-id\-field\fR earlier names take precedence over later ones. .HP \fB\-\-adjust\-ref\-field\fR .HP \fB\-\-adjust\-alt\-field\fR .HP \fB\-\-adjust\-a1\-field\fR .HP \fB\-\-adjust\-test\-field\fR .HP \fB\-\-adjust\-p\-field\fR .TP \fB\-\-ci\fR : Report confidence ratios for odds ratios/betas. .TP \fB\-\-pfilter\fR : Filter out assoc. test results with higher p\-values. .TP \fB\-\-score\-col\-nums\fR <...> : Process all the specified coefficient columns in the \fB\-\-score\fR file, identified by 1\-based indexes and/or ranges of them. .HP \fB\-\-q\-score\-range\fR [i] [j] ['header'] ['min'] : .IP Apply \fB\-\-score\fR to subset(s) of variants in the primary score list(s) based on e.g. p\-value ranges. * The first file should have range labels in the first column, p\-value .IP lower bounds in the second column, and upper bounds in the third column. Lines with too few entries, or nonnumeric values in the second or third column, are ignored. .IP * The second file should contain a variant ID and a p\-value on each line .TP (except possibly the first). Variant IDs are read from column #i and .IP p\-values are read from column #j, where i defaults to 1 and j defaults to i+1. The 'header' modifier causes the first nonempty line of this file to be skipped. .IP * By default, \fB\-\-q\-score\-range\fR errors out when a variant ID appears multiple .TP times in the data file (and is also present in the main dataset). To use .IP the minimum p\-value in this case instead, add the 'min' modifier. .TP \fB\-\-vscore\-col\-nums\fR <...> : Process all the specified coefficient columns in the \fB\-\-variant\-score\fR file, identified by 1\-based indexes and/or ranges of them. .TP \fB\-\-parallel\fR : Divide the output matrix into n pieces, and only compute the kth piece. The primary output file will have the piece number included in its name, e.g. plink2.king.13 or plink2.king.13.zst if k is 13. Concatenating these files in order will yield the full matrix of interest. (Yes, this can be done before decompression.) N.B. This generally cannot be used to directly write a symmetric square matrix. Choose square0 or triangle shape instead, and postprocess as necessary. .TP \fB\-\-memory\fR ['require'] : Set size, in MiB, of initial workspace malloc attempt. To error out instead of reducing the request size when the initial attempt fails, add the 'require' modifier. .TP \fB\-\-threads\fR : Set maximum number of compute threads. .TP \fB\-\-d\fR : Change variant/covariate range delimiter (normally '\-'). .TP \fB\-\-seed\fR : Set random number seed(s). Each value must be an integer between 0 and 4294967295 inclusive. Note that \fB\-\-threads\fR and "\-\-memory require" may also be needed to reproduce some randomized runs. .TP \fB\-\-output\-min\-p\fR

: Specify minimum p\-value to write to reports. (2.23e\-308 is useful for preventing underflow in some programs.) .TP \fB\-\-debug\fR : Use slower, more crash\-resistant logging method. .TP \fB\-\-randmem\fR : Randomize initial workspace memory (helps catch uninitialized\-memory bugs). .TP \fB\-\-warning\-errcode\fR : Return a nonzero error code to the OS when a run completes with warning(s). .TP \fB\-\-zst\-level\fR : Set the Zstd compression level (1\-22, default 3). .PP Primary methods paper: Chang CC, Chow CC, Tellier LCAM, Vattikuti S, Purcell SM, Lee JJ (2015) Second\-generation PLINK: rising to the challenge of larger and richer datasets. GigaScience, 4. .SH "SEE ALSO" The full documentation for .B PLINK v2.00a3 64-bit (9 Apr is maintained as a Texinfo manual. If the .B info and .B PLINK v2.00a3 64-bit (9 Apr programs are properly installed at your site, the command .IP .B info PLINK v2.00a3 64-bit (9 Apr .PP should give you access to the complete manual.