Overview
About
OSCA(OmicS-data-based Complex trait Analysis) is a software tool for the analysis of complex traits using multi-omics data and genetic analysis of molecular phenotypes.
Functions currently supported are:
Estimating the epigenetic (or transcriptomic) relationships between individuals from genome-wide DNA methylation (or gene expression) data – the ORM method.
Estimating the proportion of phenotypic variance for a complex trait that can be captured by all DNA methylation (or gene expression) probes – the OREML method.
Mixed linear model-based approaches to test for associations between DNA methylation (or gene expression) probes and complex traits (i.e., the MOA and MOMENT methods).
Estimating the joint "effects" of all methylation (transcription) probes on a phenotype (e.g. BMI) in a mixed linear model (analogous to BLUP). These estimated effects can be used to "predict" the phenotype in a new independent sample.
eQTL/mQTL analysis based on either a linear regression model or a mixed linear model with relevant covariates (results saved inBESD format).
vQTL analysis to test for the effects of genetic variants on the variance of a trait. Three algorithms have been implemented (i.e., the Bartlett's test, the Levene's test, and the Fligner-Killeen test).
sQTL analysis to detect variants associated with pre-mRNA splicing.
Note:although the ORM, OREML, MOA, and MOMENT methods are designed for gene expression and DNA methylation data, they can in principle be applied to any other source of omic data including microbiome, proteome and brain connectome.
Credits
Futao Zhangwrote the original version of the software code and contributed to the ORM, OREML, MOA and MOMENT methods.Ting Qicontributed to the THISTLE and MeCS methods.Zhihong Zhucontributed to the ORM and OREML methods.Hailing Fangmade several improvements in the sQTL and xQTL modules and is currently maintaining the software.Huanwei Wangcontributed to vQTL method.Shouye LiuandZhili Zhengprovided technical support.Jian Yangsupervised the project and contributed to the methods, software development and documentation.
Questions and Help Requests
Bug reports or questions to Jian Yang (jian.yang@westlake.edu.cn), School of Life Sciences, betway必威棋牌登录
Citation
ORM, OREML, MOA and MOMENT methods:
Zhang F, Chen W, Zhu Z, Zhang Q, Nabais, MF, Qi T, Deary IJ, Wray NR, Visscher PM, McRae AF, Yang J (2019) OSCA: a tool for omic-data-based complex trait analysis.Genome Biol, 20:107PMID:31138268
vQTL analysis:
Wang H, Zhang F, Zeng J, Wu Y, Kemper KE, Xue A, Zhang M, Powell JE, Goddard ME, Wray NR, Visscher PM, McRae AF, Yang J (2019) Genotype-by-environment interactions inferred from genetic effects on phenotypic variability in the UK Biobank.Science Advances, 5(8):eaaw3538PMID:31453325
MeCS (eQTL meta-analysis):
Qi T, Wu Y, Zeng J, Zhang F, Xue A, Jiang L, Zhu Z, Kemper K, Yengo L, Zheng Z; eQTLGen Consortium, Marioni RE, Montgomery GW, Deary IJ, Wray NR, Visscher PM, McRae AF, Yang J (2018) Identifying gene targets for brain-related traits using transcriptomic and methylomic data from blood.Nature Communications, 9(1):2282PMID:29891976
THISTLE (sQTL mapping):
Qi T, Wu Y, Fang H, Zhang F, Liu S, Zeng J, Yang J (2022)Genetic control of RNA splicing and its distinct role in complex trait variation. Nature Genetics, in press.
Download
Executable Files (version 0.46)
The Linux version is available at:osca-0.46.1-linux-x86_64.zip.
The MacOS version is available at:osca-Mac-v0.45.zip.
The Windows version is available at:osca-Win-v0.45.zip.
The example files are inexample.zip.
The executable files (binary code) are release under MIT lincense.
Source code (version 0.46)
https://github.com/jianyangqt/osca
The source code is released under GPL v2.
Update log
8.Version 0.45 (17 June, 2019): released a beta version of MOMENT2.osca_v0.45_linux_static.zip
7.Version 0.44 (3 May, 2019): updated the PCA function and released the exact MOA and MOMENT approaches (more accurate but computationally less efficient than the default approximate approaches).
6.Version 0.43 (18 January, 2019): added a function to run a variance quantitative trait locus (vQTL) analysis.
6.Version 0.42 (29 November, 2018): added an efficient function to run a mixed-linear-model-based cis-eQTL analysis.
5.Version 0.41 (22 October, 2018): added logistic regression in EWAS analysis.
4.Version 0.40 (14 August, 2018): added a new method called MOMENT for EWAS analysis and some functions for Data Management.
3.Version 0.39 (14 February, 2018): added a solution if the MLM estimation process in EWAS analysis does not work.
2.Version 0.38 (8 January, 2018): added meta analysis, eQTL analysis.
1.21 July, 2017: first release.
Data Management
The DNA methylation (or gene expression) data are stored in a binary format for storage efficiency. We store the data in three separate files .oii (individual information, similar as a PLINK .fam file), .opi (probe information) and .bod (a binary file to store the DNA methylation or gene expression profiles).
To efficiently save the result of eQTL / vQTL analysis, OSCA also supports BESD format.
BOD format
# BOD format: an efficient format to store omics data
myeed.oii
F01 I01 0 0 NA F02 I01 0 0 NA F03 I02 0 0 NA ...Columns are family ID, individual ID, paternal ID, maternal ID and sex (1=male; 2=female; 0=unknown). Missing data are represented by "NA".
myeed.opi
1 probe101 924243 Gene01 + 1 probe102 939564 Gene01 - 1 probe103 1130681 Gene01 - ...Columns are chromosome, probe ID (can be the ID of an exon or a transcript for RNA-seq data), physical position, gene ID and gene orientation.
myeed.bod
DNA methylation (or gene expression) data in binary format.Please do not try to open this file with a text editor.
Note:The text below are for advanced users only. We use the first 12 bytes of the .bod file to store the descriptive information for the data. The first 2 bytes are reserved for further extension of the software. The type of data is indicated by the 3rd byte (0 for gene expression data, 1 for DNA methylation data, and 2 for any other type of data). The type of value is indicated by the 4th byte (0 for DNA methylation beta value, 1 for DNA methylation m value and 2 for any other type of value). The number of individuals and number of probes are stored as integers (4 bytes each). For example, for a DNA methylation data set (0x01) of m values (0x01) on 1,342 (0x53e) individuals and 228,694 (0x37d56) CpG sites, the first 12 bytes of the .bod file are:
0101 0000 3e05 0000 567d 0300The bytes afterwards are for the DNA methylation or gene expression data.
Make a BOD file
# compile data in binary format from text format
osca --efile myprofile.txt --methylation-beta --make-bod --out myprofile--efilereads a DNA methylation (or gene expression) data file in plain text format.
--methylation-betaindicates methylation beta values in the file.
--make-bodsaves DNA methylation (or gene expression) data in binary format.
--outsaves data ( or results) in a file.
myprofile.txt
FID IID cg00000658 cg26036652 cg00489772 ... F01 I01 0.909 0.845 0.41 ... F01 I02 0.832 0.732 0.503 ... ...This is a file with a header line that contains family ID, individual ID and names of probes. The column of family ID is optional. Please use the flag "--no-fid" for data without family ID.
osca --efile myprofile.txt --methylation-beta --make-bod --no-fid --out myprofile--no-fidindicates data without family ID.
osca --efile myprofile.txt --methylation-m --make-bod --out myprofile--methylation-mindicates DNA methylation m values in the file.
If the profile is of gene expression value.
osca --efile myprofile.txt --gene-expression --make-bod --out myprofile--gene-expressionindicates gene expression profiles in the file.
For any other type of data:
osca --efile myprofile.txt --make-bod --out myprofileA binary file also can be made from a transposed file in text format.
osca --tefile mytprofile.txt --methylation-beta --make-bod --no-fid --out myprofilemytprofile.txt
FID F01 F01 ... IID I01 I02 ... cg00000658 0.909 0.832 ... cg26036652 0.845 0.732 ... cg00489772 0.41 0.503 ... ...The row of family ID is optional. Please use the flag "--no-fid" if there is no family ID in your data.
osca --tefile mytprofile.txt --methylation-beta --make-bod --no-fid --out myprofile# update probe annotation
Probe information are sometimes not available in the original DNA methylation (or gene expression) data. These information can be updated the command below.
osca --befile myprofile --update-opi annotated.opi--befilereads a DNA methylation (or gene expression) data file in binary format.
--update-opireads a fully annotated .opi file.
Manage BOD file(s)
# make a subset
To extract a probe
osca --befile myprofile --probe cg00000658 --make-bod --out mysubprofile--probeextracts a specified probe.
To exclude a probe
osca --befile myprofile --probe-rm cg00000658 --make-bod --out mysubprofile--probe-rmexcludes a specified probe.
To extract a subset of probes
osca --befile myprofile --extract-probe probe.list --make-bod --out mysubprofile--extract-probeextracts a subset of probes.
probe.list
cg00000658 cg26036652 ...To exclude a subset of probes
osca --befile myprofile --exclude-probe probe.list --make-bod --out mysubprofile--exclude-probeexcludes a subset of probes.
To extract a subset of individuals
osca --befile myprofile --keep indi.list --make-bod --out mysubprofile--keepextracts a subset of individuals.
indi.list
F01 I01 F01 I02 ...To exclude a subset of individuals
osca --befile myprofile --remove indi.list --make-bod --out mysubprofile--removeexcludes a subset of individuals.
To extract a subset of genes
osca --befile myprofile --genes gene.list --make-bod --out mysubprofile--genesextracts a subset of genes.
gene.list
MAN1B1 EDEM2 ...To extract a gene
osca --befile myprofile --gene MAN1B1 --make-bod --out mysubprofile--geneextracts a gene.
To extract a subset of probes based on the order
osca --befile myprofile --from-probe probe0 --to-probe probe1 --make-bod --out mysubprofile--from-probespecifies the start probe.
--to-probespecifies the end probe.
To extract a subset of probes in a genomic region centred at a specific probe
osca --befile myprofile --probe cg00000658 --probe-wind 500 --make-bod --out mysubprofile--probe-winddefines a window \
To extract a subset of probes on a chromosome
osca --befile myprofile --chr 1 --make-bod --out mysubprofile--chrspecifies a chromosome to select probes.
To extract a subset of probes based on physical positions
osca --befile myprofile --from-probe-kb 100 --to-probe-kb 200 --make-bod --out mysubprofile--from-probe-kbspecifies the start physical position of the probes.
--to-probe-kbspecifies the end physical position of the probes.
# merge BOD files
osca --befile-flist mybod.flist --make-bod --out myprofile--befile-flistreads a file to get the full paths of the binary files.
mybod.flist
path1/my_bod1 path2/my_bod2 ...# Quality control
To remvoe probes with low variance
osca --befile myprofile --sd-min 0.02 --make-bod --out newprofile--sd-minremoves the probes with the standard deviation smaller than a specified threshold.
To remove probes with a missing rate threshold
osca --befile myprofile --missing-ratio-probe 0.01 --make-bod --out newprofile--missing-ratio-probespecifies a missing proportion threshold to remove probes.
# Quality control of methyaltion data
To remove constitutively methylated/unmethylated probes
osca --befile myprofile --upper-beta 0.8 --lower-beta 0.2 --make-bod --out newprofile--upper-betaremoves the DNA methylation probes with the mean beta value larger than a specified threshold.
--lower-betaremoves the DNA methylation probes with the mean beta value smaller than a specified threshold.
To QC with detection p-values
osca --befile myprofile --detection-pval-file dpval.txt --dpval-mth 0 --dpval-thresh 0.05 --make-bod --out newprofile--detection-pval-filereads a file that contains DNA methylation detection p-values.
--dpval-mthspecifies a method to do quality control with the detection p-values, 0 for removing the probes with one or more detection p-values violating a threshold (default as 0.05), and 1 for dropping the samples violating a proportion threshold (default as 1%) and simultaneously dropping the probes violating a proportion threshold (default as 1%).
--dpval-threshspecifies a threshold of detection p-value.
osca --befile myprofile --detection-pval-file dpval.txt --dpval-mth 1 --ratio-probe 0.01 --ratio-sample 0.01 --dpval-thresh 0.05 --make-bod --out newprofile--ratio-probespecifies a proportion threshold to remove probes.
--ratio-samplespecifies a proportion threshold to remove individuals.
the file "dpval.txt" is in the same format with a transposed profile file. Option "--no-fid" is also valid here.
# adjust the covariates for each probe
osca --befile myprofile --covar my.cov --qcovar my.qcov --adj-probe --make-bod --out newprofile--adj-probeadjusts the covariates for each probe.
# standardize each probe
osca --befile myprofile --std-probe --make-bod --out newprofile--std-probestandardizes each probe.
osca --befile myprofile --rint-probe --make-bod --out newprofile--rint-probenormalizes each probe by Rank-based Inverse Normal Transformation.
# other options
Transform methylation beta value to methylation m value and vice versa
osca --befile myprofile --m2beta --make-bod --out newprofile--m2betacalculates the methylation beta value from the methylation m value.
osca --befile myprofile --beta2m --make-bod --out newprofile--beta2mcalculates the methylation m value from the methylation beta value.
Query a BOD file
# make a text format file
osca --befile myprofile --make-efile --out myprofile.txt--make-efilesaves the DNA methylation (or gene expression) data in text format.
osca --befile myprofile --make-tefile --out mytprofile.txt--make-tefilesaves the DNA methylation (or gene expression) data in transposed text format.
NOTE: The options inManage BOD file(s)can also be applied.
# Calculate the variance and the mean of probes
osca --befile myprofile --get-variance --get-mean --out newprofile--get-variancecalculates the variance of each probe.
--get-meancalculates the mean of each probe.
newprofile.var.txt
cg00000957 0.000812127 cg00001349 0.00560701 ...newprofile.mean.txt
cg00000957 0.901574 cg00001349 0.860279 ...BESD format
Results from eQTL analysis will be saved in SMR BESD formatsee SMR website for more information.To save the eQTL results more efficiently, we extended the BESD format.
Note:The texts below are for advanced users only.
BESD dense format 1: the first 64 Bytes are reserved for the descriptive information which starts with 0x00000005. The data onwards are a vector of effect sizes followed by a vector of SEs of each probe across all the snps. The effect sizes and SEs are saved in single float-precision (4B).
BESD dense format 2(only supported in OSCA): the first 64 Bytes are reserved for the descriptive information which starts with 0x00000004. The data onwards are the effect sizes followed by the SEs of each SNP across all the probes. The effect sizes and SEs are saved in single float-precision (4B).
BESD sparse format: the first 64 Bytes are reserved for the descriptive information which starts with 0x00000001. The effect sizes and the SEs as saved inthe CSC format
Query a BESD file
This feasuer is memory efficent even to query with huge dense BESD file.
# Command line options for SNPs
To query the eQTL resutls for a single SNP, we could use this command
osca --beqtl-summary myeqtl --query 5.0e-8 --snp rs123 --out myquery--querysaves in text format a subset of the eQTL summary dataset based on the specified eQTL p-value threshold. The default value is 5.0e-8.
--snpspecifies a single SNP.
myquery.txt
SNP Chr BP A1 A2 Freq Probe Probe_Chr Probe_bp Gene Orientation b se p rs01 1 1001 A G 0.23 cg01 1 1101 gene1 + -0.033 0.006 3.8e-08 rs01 1 1001 A G 0.06 cg02 1 1201 gene2 - 0.043 0.007 8.1e-10 ......To query the eQTL resutls excluding a single SNP, we could use this command
osca --beqtl-summary myeqtl --query 5.0e-8 --snp-rm rs123 --out myquery--snp-rmspecifies a single SNP to exclude.
To query the eQTL resutls extracting a subset of SNPs , we could use this command
osca --beqtl-summary myeqtl --query 5.0e-8 --extract-snp mysnp.list --out myquery--extract-snpextracts a subset of SNPs for analysis.
To query the eQTL resutls excluding a subset of SNPs , we could use this command
osca --beqtl-summary myeqtl --query 5.0e-8 --exclude-snp mysnp.list --out myquery--exclude-snpexcludes a subset of SNPs from analysis.
To query eQTL resutls for a range of SNPs in a genomic region
osca --beqtl-summary myeqtl --query 5.0e-8 --from-snp rs123 --to-snp rs456 --out myquery--from-snpspecifies the start SNP.
--to-snpspecifies the end SNP.
NOTE: All SNPs should be on the same chromosome.
To query eQTL results for all SNP on a chromosome
osca --beqtl-summary myeqtl --query 5.0e-8 --snp-chr 1--snp-chrspecifies a chromosome to select SNPs.
NOTE: The probes in the result could be on the other chromosomes if there aretrans-eQTLs.
To query SNPs based on physical positions
osca --beqtl-summary myeqtl --query 5.0e-8 --snp-chr 1 --from-snp-kb 100 --to-snp-kb 200 --out myquery--from-snp-kbspecifies the start physical position of the region.
--to-snp-kbspecifies the end physical position of the region.
NOTE: You will need to specify a chromosome (using the '--snp-chr' option) when using this option.
To query based on a flanking region of a SNP
osca --beqtl-summary myeqtl --query 5.0e-8 --snp rs123 --snp-wind 50 --out myquery--snp-winddefines a window centred on a specified SNP.
# Command line options for probes
To query based on a single probe
osca --beqtl-summary myeqtl --query 5.0e-8 --probe cg123 --out myquery--probespecifies a single probe.
To query excluding a single probe
osca --beqtl-summary myeqtl --query 5.0e-8 --probe-rm cg123 --out myquery--probe-rmspecifies a single probe to exclude.
To query the eQTL resutls extracting a subset of probes , we could use this command
osca --beqtl-summary myeqtl --query 5.0e-8 --extract-probe myprobe.list --out myquery--extract-probeextracts a subset of probes for analysis.
To query the eQTL resutls excluding a subset of probes , we could use this command
osca --beqtl-summary myeqtl --query 5.0e-8 --exclude-probe myprobe.list --out myquery--exclude-probeexcludes a subset of probes from analysis.
To query based on a range of probes
osca --beqtl-summary myeqtl --query 5.0e-8 --from-probe cg123 --to-probe cg456 --out myquery--from-probespecifies the start probe.
--to-probespecifies the end probe.
NOTE : All probes should be on the same chromosome.
To query based on a chromosome
osca --beqtl-summary myeqtl --query 5.0e-8 --probe-chr 1--probe-chrspecifies a chromosome to select probes.
NOTE: The SNPs in the result could be on the other chromosomes if there aretrans-eQTLs.
To query based on physical positions of the probes
osca --beqtl-summary myeqtl --query 5.0e-8 --probe-chr 1 --from-probe-kb 1000 --to-probe-kb 2000 --out myquery--from-probe-kbspecifies the start physical position of the probes.
--to-probe-kbspecifies the end physical position of the probes.
NOTE: You will need to specify a chromosome (using the '--probe-chr' option) when using this option.
To query based on a flanking region of a probe
osca --beqtl-summary myeqtl --query 5.0e-8 --probe cg123 --probe-wind 1000 --out myquery--probe-winddefines a window centred on a specified probe.
To query based on a gene
osca --beqtl-summary myeqtl --query 5.0e-8 --gene gene1 --out myquery--genespecifies a single gene to select probes.
# Command line option for cis-region
osca --beqtl-summary myeqtl --query 5.0e-8 --probe cg123 --cis-wind 2000 --out myquery# File-list options
To query based on a list of SNPs
osca --beqtl-summary myeqtl --extract-snp snp.list --query 5.0e-8 --out myqueryTo query based on a list of probes
osca --beqtl-summary myeqtl --extract-probe probe.list --query 5.0e-8 --out myqueryTo qurey based on a list of genes
osca --beqtl-summary myeqtl --genes gene.list --query 5.0e-8 --out myquery--genesextracts a subset of probes which tag the genes in the list.
gene.list
gene1 gene2 gene3 ...Manage a BESD file
# make a subset of data
All the data management options inQuery a BESD filecan be here.
for example, To extract a subset of SNPs and/or probes
osca --beqtl-summary myeqtl --extract-snp mysnp.list --extract-probe myprobe.list --make-besd --out mybesd# tansform to SMR compatible format
osca --beqtl-summary myeqtl --make-besd --to-smr --out mybesd--to-smrtranforms BESD file to SMR compatible format.
# shrink a BESD file
To remove probes without any value across all the SNPs and SNPs without any value across all the probes.
osca --beqtl-summary myeqtl --make-besd --besd-shrink --out mybesd--besd-shrinkremoves probes and SNPs that have no value.
ORM
estimate ORM
osca --befile myprofile --make-orm --out myormosca --befile myprofile --make-orm-bin --out myorm--make-ormor--make-orm-binestimates the omics relationship matrix (ORM) between pairs of individuals from a set of probes and save the lower triangle elements of ORM to binary files. This format is compatible with theGRMin theGCTAsoftware.
osca --befile myprofile --make-orm-gz --out myorm--make-orm-gzestimates the ORM and save the lower triangle elements of ORM to compressed plain text files.
osca --befile myprofile --make-orm --orm-alg 1 --out myorm--orm-algspecifies the algorithm to estimate the ORM. 1 for standardized data of each probe, 2 for centred data of each probe and 3 for iteratively standardizing probes and individuals. The default option is 1.
Note that although we describe the options above using DNA methylation and gene expression data, all the options can be applied to any other source of omics data mentioned above.
Principal Component Analysis
osca --orm myorm --pca 20 --out mypca--ormreads the ORM binary files.
--pcaconducts principal component analysis and saves the first n (default as 20) PCs.
Principal components can also be directly computed from data inBOD format
osca --befile myprofile --pca 20 --out mypcamypca.eigenval
122.014 95.3064 70.8055 ...mypca.eigenvec
R06C01 R06C01 0.012637 -0.00328532 0.0263842 -0.00595246 R05C02 R05C02 0.0106692 -0.0184545 -0.0159826 0.013951 R04C02 R04C02 0.0024727 -0.0118185 -0.0256503 -0.0106235 ...Users can also manipulate the ORM in the analysis.
osca --orm myorm --keep indi.list --pca 20 --out mypcaosca --orm myorm --remove indi.list --pca 20 --out mypcaosca --orm myorm --orm-cutoff 0.05 --pca 20 --out mypca--orm-cutoffremoves one of a pair of individuals with estimated omics relationships larger than the specified cut-off value.
osca --multi-orm myorm.flist --pca 20 --out mypca--multi-ormreads multiple ORMs in binary format.
OREML
osca --reml --pheno my.phen --out myreml--remlperforms REML (restricted maximum likelihood) analysis. This option is usually followed by the option--orm(one ORM) or--merge-orm(multiple ORMs) to estimate the variance explained by the probes that were used to estimate the omics relationship matrix.
--phenoreads phenotype data from a plain text file. Missing value should be represented by "NA".
my.phen
R06C01 R06C01 32.6332 R05C02 R05C02 23.9411 R04C02 R04C02 29.7441 ...NOTE:current version only supports single trait analysis in one run.
osca --reml --orm myorm --pheno my.phen --out myremlosca --reml --orm myorm --pheno my.phen --keep indi.list --out myremlosca --reml --orm myorm --pheno my.phen --orm-cutoff 0.05 --out myremlWith multiple ORMs
osca --reml --multi-orm myorm.flist --pheno my.phen --out myremlosca --reml --multi-orm myorm.flist --pheno my.phen --reml-alg 0 --out myreml--reml-algspecifies the algorithm to do REML iterations, 0 for average information (AI), 1 for Fisher-scoring and 2 for EM. The default option is 0, i.e. AI-REML, if this option is not specified.
osca --reml --multi-orm myorm.flist --pheno my.phen --reml-maxit 100 --out myreml--reml-maxitspecifies the maximum number of iterations. The default number is 100 if this option is not specified.
osca --reml --orm myorm --pheno my.phen --covar my.covar --out myreml--covarreads discrete covariates from a plain text file.
my.covar
R06C01 R06C01 F 0 R05C02 R05C02 F 1 R04C02 R04C02 M 1 ...osca --reml --orm myorm --pheno my.phen --qcovar my.qcovar --out myreml--qcovarreads quantitative covariates from a plain text file.
my.qcovar
R06C01 R06C01 25 R05C02 R05C02 16 R04C02 R04C02 30 ...osca --reml --orm myorm --pheno my.phen --reml-est-fix --out myreml--reml-est-fixdisplays the estimates of fixed effects on the screen.
osca --reml --orm myorm --pheno my.phen --reml-no-lrt --out myreml--reml-no-lrtturns off the LRT.
EWAS
MOMENT
Multi-componentMLM-based associationexcluding thetarget
# MOMENT (approximate approach)
osca --moment --befile myprofile --pheno my.phen --out my--momentruns an approximate MOMENT analysis, i.e., a multi-component MLM based association test excluding the target probe (the probe to be tested for association) and the probes in its flanking region from the ORM (the size of flanking region can be modified by--moment-wind). The probes are classified into two groups by the association statistics (in descending order) from a linear regression analysis. This is an approximate approach in which each target probe is tested using the estimated variance components under the null model. The variance components are only re-estimated when one or more probes in the first group (the group in which the probes have larger linear regression association statistics than those in the other group) are excluded from the ORMs. The results will be saved in a plain text file (*.moment).
osca --moment --befile myprofile --pheno my.phen --moment-wind 100 --out my--moment-windspecifies a flanking region to exclude probes from the ORM. The default value is 100 Kb (i.e. a 100 Kb region centred at the probe to be tested).
my.mlma
Chr Probe bp Gene Orientation b se p 1 cg00003287 201346149 TNNT2 - -0.156 0.597 0.794 1 cg00008647 207082900 IL24 + 0.032 0.354 0.926 1 cg00009292 50882082 DMRTA2 - 0.120 1.182 0.919 ...This is a text file with headers. Columns are chromosome, probe, probe BP, gene, orientation, effect size, standard error and p-value.
# MOMENT2 ( beta version)
In MOMENT, the probes are divided into two groups by their association p-values generated by linear regression (note: the default threshold p-value is 0.05 / m with m being the total number of probes) and fitted as two random-effect components in the model to control for unobserved confounding effects (Zhang et al. 2019 Genome Biology). When the number of probes in the first group is too large, the analysis will become slow because the program needs to re-estimate the variance components whenever one or more probes are removed from the first group (to avoid proximity contamination; seeZhang et al.for details). This may also cause convergence problem because of too much variation explained by the first random-effect component. In MOMENT2, a stepwise selection procedure is implemented to reduce the number of probes in the first group. Simulation shows that MOMENT2 has approximately the same level of false positive rate but slightly higher power compared to MOMENT. MOMENT2 is also expected to be faster than MOMENT especially when the number of associated probes detected by linear regression is large.
osca --moment2-beta --befile myprofile --pheno my.phen --out my--moment2-betaruns a MOMENT2 analysis.
When the chromosome or the position information is unknown for some kinds of data such as Microbiome, MRI (Magnetic resonance imaging), etc., MOMENT / MOMENT2 will automatically switch window-based method to correlation-based method to exclude the target probe and its highly correlated probes. The correlation-based method can be also specified to run by--moment-corand the default correlation R-squared threshold is 0.6 which can be modified by--cor-r2.
osca --moment2-beta --befile myprofile --pheno my.phen --moment-cor --cor-r2 0.6 --out my--moment-corruns a correlation-based method to exclude the target probe and its highly correlated probes.
--cor-r2specifies a correlation r2 to select highly correlated probes with the target probe. The default value is 0.6.
# MOMENT (exact approach)
osca --moment-exact --befile myprofile --pheno my.phen --out my--moment-exactruns an exact MOMENT test. This is similar to the method above (window size can also be modified by--moment-wind) but the variance components are re-estimated for each target probe.
NOTE:Sometimes the MLM estimation process in the EWAS analysis does not work (the REML iteration cannot converge or an estimate hits the boundaries of parameter space) especially when the sample size is small. In such case MLM will be terminated.
MOA
MLM-basedomicassociation
# MOA (approximate approach)
osca --moa --befile myprofile --pheno my.phen --out myIf you have already computed the ORM
osca --moa --befile myprofile --pheno my.phen --orm myorm --out my--moarun an approximate MOA analysis, i.e., an MLM based association test including the target probe in the ORM. This an approximate approach in which each target probe is tested using the variance components estimated under the null model. The results will be saved in a plain text file (*.moa).
# MOA (exact approach)
osca --moa-exact --befile myprofile --pheno my.phen --out my--moa-exactruns an exact MOA test.
Multi-threading computation
osca --moa-exact --befile myprofile --pheno my.phen --task-num 1000 --task-id 1 --thread-num 10 --out myLinear Regression
osca --befile myprofile --pheno my.phen --linear --out myosca --befile myprofile --pheno my.phen --qcovar my.qcovar --covar my.covar --linear --out my--linearsaves linear regression statistics to a plain text file.
my.linear
probeChr ProbeID Probe_bp Gene Orientation BETA SE P NMISS 1 cg00003287 201346149 TNNT2 - -0.0594 0.531 9.11e-01 1337 1 cg00008647 207082900 IL24 + 0.5003 0.263 5.72e-02 1337 1 cg00009292 50882082 DMRTA2 - 1.0512 1.026 3.06e-01 1337 ...This is a text file with headers. Columns are chromosome, probe, probe BP, gene, orientation, effect size, standard error, p-value and number of non-missing individuals.
Fast Linear Regression
The options above fit the covariates in the model as \(y=xb+C\beta+e\). This is equivalent to \(y^{'}=y-C\hat\beta\), \(x^{'}=x-C\hat\beta\), \(y^{'}=x^{'}b\). This equivalent transformation avoids the repeated computing of the inverse of a \((p +1)\times(p + 1)\) matrix where \(p\)is the number of covariates.
osca --linear --befile myprofile --pheno my.phen --qcovar my.qcovar --covar my.covar --fast-linear --out my--fast-linearruns a fast linear regression analysis.
Logistic Regression
osca --befile myprofile --pheno my.phen --logistic --out myosca --befile myprofile --pheno my.phen --qcovar my.qcovar --covar my.covar --logistic --out my--logisticoutputs logistic regression analysis result to a plain text file.
my.logistic
probeChr ProbeID Probe_bp Gene Orientation OR SE P NMISS 1 cg00297950 110282525 GSTM3 - 0.010953 1.30438 5.386424e-04 1318 1 cg00299820 11943154 NPPB - 0.0085006 2.77659 8.596523e-02 1318 1 cg00305285 1017115 RNF223 - 8.22097 2.99934 4.824397e-01 1318 ...This is a text file with headers. Columns are chromosome, probe, probe BP, gene, orientation, odd ratio, standard error, p-value and number of non-missing individuals.
NOTE:It is allowed to use characters, strings, or numbers as phenotype values in logistic regression but "NA" or "na" will be recognised as a missing value.
EWAS simulation
The phenotypes are simulated based on a set of real DNA methylation (or gene expression) data and a simple model y= sum(xibi)+ε, where y is a vector of phenotypes, xiis a vector of raw DNA methylation (or gene expression) profile or standardized profile of the i-th "causal" probe, biis the effect of the i-th causal probe and ε ~ N(0,var( sum(xibi))(1/h2-1) ) is a vector of residual effect.
osca --simu-qt --simu-rsq 0.1 --befile myprofile --simu-causal-loci mycausal.list --out mypheno--simu-qtsmulates a quantitative trait.
--simu-rsqspecifies the proportion of variance in phenotype explained by the causal probes. The default value is 0.1.
--simu-causal-locireads a list of probes as causal probes. If the effect sizes are not specified in the file, they will be generated from a standard normal distribution.
mycausal.list
cg04584301 1.55182 cg04839274 -0.106226 cg16648571 0.0257417 ...This is a text file with no headers. Columns are probe ID and effect size.
osca --simu-qt --simu-hsq 0.1 --simu-eff-mod 0 --befile myprofile --simu-causal-loci mycausal.list --out mypheno--simu-eff-modspecifies whether or not to standardize the causal probe, 0 for standardized profile and 1 for raw profile. The default value is 0.
osca --simu-cc 100 300 --simu-hsq 0.1 --simu-k 0.1 --befile myprofile --simu-causal-loci mycausal.list --out mypheno--simu-ccsimulates a case-control trait and specifies the number of cases and the number of controls.
--simu-kspecifies the disease prevalence. The default value is 0.1 if this option is not specified.
Prediction Analysis
osca --reml --orm myorm --pheno my.phen --reml-pred-rand --out myblp--reml-pred-randpredicts the random effects by the BLUP (best linear unbiased prediction) method. This option is to estimate the aggregated effect of all the probes (used to compute the ORM) to the phenotype of an individual. The aggregated omics effects of all the individuals will be saved in a plain text file*.indi.blp.
myblp.indi.blp
R06C01 R06C0 1.02275 0.07065 1.05692 1.05692 R05C02 R05C02 0.18653 0.27059 0.86650 0.86650 R04C02 R04C02 -0.1982 -0.1673 -0.9209 -0.9209 ...This is a text file with no headers. Columns are family ID, individual ID, an intermediate variable, the aggregated omics effect, another intermediate variable and the residual effect.
osca --befile myprofile --blup-probe myblp.indi.blp --out myblp--blup-probecalculates the BLUP solutions for the probe effects.
myblp.probe.blp
cg04584301 -0.000654646 cg04839274 0.000602484 cg16648571 -4.70356e-05 ...This is a text file with no headers. Columns are probe ID and BLUP of the probe effect.
osca --befile myprofile --score myblp.probe.blp --out myscoreosca --befile myprofile --score myblp.probe.blp 1 2 --out myscore--scorereads score files for probes and generates predicted omics profiles for individuals. (Note that this option largely follows the--scoreoption in PLINK.) It allows users to specify the column numbers for probe ID and score (the default values are 1 and 2 as shown in the example above).
myscore.profile
FID IID PHENO CNT SCORE 131000028 422572 -9 20000 -7.269198e-07 131000031 243421 -9 20000 9.322096e-06 131000179 338728 -9 20000 -1.250443e-05 ...This is a text file with headers. Columns are family ID, individual ID, phenotype, Number of non-missing probes and score.
For example In the score file: cg04584301 -0.065 cg04839274 0.060 cg16648571 -1.03 In the DNA methylation data: FID IID cg04584301 cg04839274 cg16648571 R05C02 R05C02 0.18653 0.27059 0.86650 The score should be: ( 0.18653*(-0.065) + 0.27059*0.060 + (-1.03)*0.86650 ) / 3 = (-0.8883841) / 3 = -0.296osca --befile myprofile --score myblp.probe.blp --score-has-header --out myscore--score-has-headerindicates probe score file has headers.
eQTL/mQTL Analysis
This is a module in OSCA to perform a standard eQTL mapping analysis. We describe the module for eQTL mapping but it can be applied to genetic association analysis for all kinds of molecular traits (e.g. DNA methylation, histone modification, metabolites and Microbiome)
Linear-regression-based
# Basic option
osca --eqtl --bfile mydata --befile myprofile --out myeqtl--eqtlenables the eQTL mapping analysis.
--bfilereads individual-level SNP genotype data (in PLINK binary format), i.e. .bed, .bim, and .fam files.
By default, the results will be saved inBESD format.
# Multi-threading computation
osca --eqtl --bfile mydata --befile myprofile --task-num 100 --task-id 1 --thread-num 10 --out myeqtl--task-numspecifies the total number of tasks to partition the computation by probes.
--task-idspecifies the task IDs form 1 to the total task number.
--thread-numspecifies the number of threads for parallel computing. The default value is 1.
# Using other data management options
osca --eqtl --bfile mydata --befile myprofile --extract-snp mysnp.list --maf 0.01 --task-num 100 --task-id 1 --thread-num 10 --out myeqtl--maffilters SNPs based on the specified MAF threshold.
# Saving the result inSMR BESD format.
osca --eqtl --bfile mydata --befile myprofile --to-smr --task-num 100 --task-id 1 --thread-num 10 --out myeqtl--to-smrsaves the result in SMR compatible format.
# Fitting covariates
osca --eqtl --bfile mydata --befile myprofile --covar mycovar --qcovar myqcovar --task-num 100 --task-id 1 --thread-num 10 --out myeqtl# runing cis-eQTL analysis
osca --eqtl --bfile mydata --befile myprofile --cis --task-num 1000 --task-id 1 --thread-num 10 --out myeqtlosca --eqtl --bfile mydata --befile myprofile --cis --cis-wind 2000 --task-num 100 --task-id 1 --thread-num 10 --out myeqtl--cisruns a cis-eQTL analysis.
--cis-windspecifies a window (in Kb unit) to store all the SNPs within the window of the probe in either direction. The default value is 2000Kb.
Mixed-linear-model-based
# runing cis-eQTL analysis
osca --eqtl --bfile mydata --befile myprofile --mlm --cis --task-num 100 --task-id 1 --thread-num 10 --out myeqtl--mlmruns a mixed-linear-model-based eQTL analysis on mixed linear model.
osca --eqtl --bfile mydata --befile myprofile --mlm --cis --grm mydata --task-num 100 --task-id 1 --thread-num 10 --out myeqtl--grmreads the GRM generated byGCTA--make-grmoption.
vQTL Analysis
This is a module in OSCA to perform a variance quantitative trait locus (vQTL) mapping analysis and save the results inBESD format. We have provided 4 methods to run vQTL analysis, namelyBartlett’s test,Levene’s testwith mean,Levene’s testwith median, andFligner-Killeen test. In the output file, apart from the vQTL statistic and p-value, we also provide the vQTL effect (i.e., the effect of a SNP on differences in phenotype variance among genotype groups), computed from z-statistic using the method described in(Zhu et al. 2016 Nature Genetics). Note that the vQTL z-statistic can be computed from p-value and the sign of the slope of regressing phenotype variance against genotypes.
# Basic option
osca --vqtl --bfile mydata --pheno mypheno --out myvqtl--vqtlto turn a vQTL analysis.
--bfileto input SNP genotype data.
--phenoto input phenotype data (seeOREMLfor the input format).
myvqtl.vqtl
Chr SNP bp statistic df beta se P NMISS 21 rs144022851 14589985 2.98808 1 -0.246533 0.142619 8.387947e-02 1319 21 rs146286292 14592960 2.99122 1 -0.244279 0.141241 8.371710e-02 1319 21 rs2847443 14595264 2.40285 1 -0.209187 0.134949 1.211141e-01 1319 21 rs192179023 14600255 0.00383907 1 0.0357931 0.577678 9.505945e-01 1319 21 rs9984084 14602180 5.21829 2 0.0696413 0.0389252 7.359741e-02 1319 ...This is a text file with headers. Columns are chromosome, SNP, SNP BP, the statistic, degree of freedom, effect size, standard error, p-value and number of non-missing individuals.
# Specifying the method to test variance heterogeneity
osca --vqtl --bfile mydata --pheno mypheno --vqtl-mtd 1 --out myvqtl--vqtl-mtdto specify the method to test variance heterogeneity. 0 for Bartlett’s test, 1 for Levene’s test with mean, 2 for Levene’s test with median, 3 for Fligner-Killeen test. The default option is 0.
myvqtl.vqtl
Chr SNP bp F-statistic df1 df2 beta se P NMISS 21 rs144022851 14589985 2.03194 1 1317 -0.203255 0.142671 0.154261 1319 21 rs188453282 14591213 1.48959 1 1317 1.21945 0.999626 0.222499 1319 21 rs146286292 14592960 1.84845 1 1317 -0.192008 0.141303 0.174196 1319 21 rs2847443 14595264 1.83381 1 1317 -0.182687 0.134978 0.17591 1319 21 rs192179023 14600255 0.00102531 1 1317 0.018494 0.577679 0.974461 1319 21 rs9984084 14602180 1.73809 2 1316 0.0526684 0.0389454 0.176259 1319 ...This is a text file with headers. Columns are chromosome, SNP, SNP BP, F-statistic, K-1 degrees of freedom, N-K degrees of freedom, effect size, standard error, p-value and number of non-missing individuals.
# Including data management options
osca --vqtl --bfile mydata --pheno mypheno --extract-snp mysnp.list --maf 0.01 --out myvqtl# Running vQTL analysis for molecular phenotypes
osca --vqtl --bfile mydata --befile myprofile --out myvqtl--befileto input molecular phenotypes (e.g. DNA methylation or gene expression measures inBOD format).
# Running vQTL analysis for molecular phenotypes in cis-regions
osca --vqtl --bfile mydata --befile myprofile --cis-wind 2000 --out myvqtl--cis-windto define a window centred around the probe to select SNPs for the vQTL analysis. The default value is 2000Kb.
Notethat vQTL results for molecular traits will be saved inBESD formatwhich can be tansformed to text format usingOSCA query.
# Multi-threading computation
osca --vqtl --bfile mydata --befile myprofile --cis-wind 2000 --task-num 1000 --task-id 1 --thread-num 10 --out myvqtlCitation
Huanwei Wang, Futao Zhang, Jian Zeng, Yang Wu, Kathryn E. Kemper, Angli Xue, Min Zhang, Joseph E. Powell, Michael E. Goddard, Naomi R. Wray, Peter M. Visscher, Allan F. McRae, Jian Yang (2019) Genotype-by-environment interactions inferred from genetic effects on phenotypic variability in the UK Biobank.bioRxiv 519538; doi: https://doi.org/10.1101/519538
The vQTL summary data for 13 UKB traits are available atData Resource
THISTLE
THISTLE (testing for heterogeneity between isoform-eQTL effects) is a transcript-based method that uses either individual-level genotype and RNA-seq data or summary-level isoform-eQTL data for splicing QTL (sQTL) mapping. It has been shown by simulation that the THISTLE test-statistic is well calibrated under the null (i.e., no sQTL effect) in either the presence or absence of eQTL effect. In both simulation and real data analysis, THISTLE shows higher statistical power and computational efficiency than the competing transcript-based sQTL method.
# THISTLE cis-sQTL analysis using individual-level genotype and RNA-seq data
osca --sqtl --bfile geno_data \ --befile isoform_data \ --qcovar qcov_data.txt \ --covar cov_data.txt \ --bed gencode.v19.annotation.gtf.ensembl.bed \ --maf 0.01 \ --call 0.85 \ --thread-num 2 \ --task-num 10 \ --task-id 1 \ --to-smr \ --out outfile--sqtlperforms THISTLE sQTL analysis.
--bfilereads individual-level SNP genotype data (inPLINK binary format), i.e., .bed, .bim, and .fam files.
--befilereads individual-level isoform expression data (inBOD format), i.e., .oii, .opi, and .bod files.
--qcovarreads quantitative covariates from a plain text file (see below for an example).
qcov_data.txt
R06C01 R06C01 25 R05C02 R05C02 16 R04C02 R04C02 30 ...--covarreads discrete covariates from a plain text file (see below for an example).
cov_data.txt
R06C01 R06C01 F 0 R05C02 R05C02 F 1 R04C02 R04C02 M 1 ...--bedreads a gene annotation file in BED format (see below for an example).
gencode.v19.annotation.gtf.ensembl.bed
1 11869 14412 ENSG00000223972 + DDX11L1 1 14363 29806 ENSG00000227232 - WASH7P 1 29554 31109 ENSG00000243485 + MIR1302-11 ...This file has no header.
--to-smrsaves the sQTL result in SMR compatible format. By default, the sQTL summary statistics will be saved inSMR BESD format.
--no-isoform-eQTLturns off the output of the isoform-eQTL summary statistics. By default, THISTLE will output the isoform-eQTL summary statistics for each significant sQTL, and we have provided an R script below to visualize the isoform-eQTL effects (please click this link to download theR script). We can use this flag to mute output of the isoform-eQTLs.
--mafremoves SNPs based on a minor allele frequency (MAF) threshold.
--callremoves SNPs based on a SNP calling rate threshold.
--thread-numspecifies the number of threads for parallel computing. The default value is 1.
--task-numspecifies the total number of tasks to partition the computation by genes.
--task-idspecifies the task IDs form 1 to the total task number.
# THISTLE trans-sQTL analysis using individual-level genotype and RNA-seq data
osca --sqtl --bfile geno_data \ --trans –trans-wind 5000 \ --befile isoform_data \ --qcovar qcov_data.txt \ --covar cov_data.txt \ --bed gencode.v19.annotation.gtf.ensembl.bed \ --maf 0.01 \ --call 0.85 \ --thread-num 2 \ --task-num 10 \ --task-id 1 \ --to-smr \ --out outfile--transperforms a THISTLE trans-sQTL analysis.
--trans-windspecifies the size of a trans-region (in Kb units). The default value is 5000 Kb.
# THISTLE cis-sQTL analysis using summary-level isoform-eQTL data
osca --sqtl --beqtl-summary isoform_eQTL_data \ --bed gencode.v19.annotation.gtf.ensembl.bed \ --thread-num 2 \ --task-num 10 \ --task-id 1 \ --to-smr \ --out outfile--beqtl-summaryreads summary-level isoform-eQTL data (inSMR BESD format), i.e., .epi, .esi, and .besd files.
# A permutation-based procedure for multiple testing correction for cis-sQTLs
osca --sqtl --bfile geno_data \ --befile isoform_data \ --qcovar qcov_data.txt \ --covar cov_data.txt \ --bed gencode.v19.annotation.gtf.ensembl.bed \ --maf 0.01 \ --call 0.85 \ --thread-num 2 \ --task-num 10 \ --task-id 1 \ --to-smr \ --permutation \ --permu-times 1000 \ --out outfile--permutationperforms permutations of the isoform-level transcriptional abundance phenotype across individuals.
--permu-timesspecifies the permutation times. The default number is 1000 if this option is not specified.
Citation
Qi T, Wu Y, Fang H, Zhang F, Liu S, Zeng J, Yang J (2022)Genetic control of RNA splicing and its distinct role in complex trait variation. Nature Genetics, in press.
BrainMeta sQTL & eQTL portal
An online tool to query or visualize the BrainMeta sQTLs and eQTLs is available at theBrainMeta data portal.
Meta-analysis
Meta-analysis
# Meta for GWAS summary data without sample overlap
osca --gwas-flist mygwas.flist --meta --out mymeta--metaimplements the conventional inverse-variance-weighted meta-analysis meta-analysis assuming all the cohorts are independent. Pleae refer tode Bakker PI et al.2008 Hum Mol Genetfor the details.
--gwas-flistreads a file to get file paths of the GWAS summary data.
mygwas.flist
Height.01.COJO Height.02.COJO Height.03.COJO ...This file has no header.
The input format of the GWAS summary data follows that for GCTA-COJO analysis (http://cnsgenomics.com/software/gcta/#COJO).
Height.01.COJO
SNP A1 A2 freq b se p n rs1001 A G 0.8493 0.0024 0.0055 0.6653 129850 rs1002 C G 0.03606 0.0034 0.0115 0.7659 129799 rs1003 A C 0.5128 0.045 0.038 0.2319 129830 ......Columns are SNP, the effect (coded) allele, the other allele, frequency of the effect allele, effect size, standard error, p-value and sample size. The headers are not keywords and will be omitted by the program.Important: “A1” needs to be the effect allele with “A2” being the other allele and “freq” needs to be the frequency of “A1”.NOTE: For a case-control study, the effect size should be log(odds ratio) with its corresponding standard error.
# Meta for eQTL summary data
osca --besd-flist mybesd.flist --meta --out mymeta--besd-flistreads a file to get the full paths of the BESD files.
mybesd.flist
path1/my_besd1 path2/my_besd2 path3/my_besd3 ...This file has no header. The eQTL summary data should be inBESD format.
MeCS
# MeCS for eQTL summary data in correlated samples
MeCS is a method that only requires summary-level cis-eQTL data to perform a meta-analysis of cis-eQTLs from multiple cohorts (or tissues) with sample overlaps. It estimates the proportion of sample overlap from null SNPs in the cis-regions and meta-analysed the eQTL effects using a generalized least squares approach. The method can be applied to data from genetic studies of molecular phenotypes (e.g. DNA methylation and histone modification).
NOTE:Only the information in the cis-region would be used.
osca --besd-flist mybesd.flist --mecs --out mymecsOptions Reference
| --befile | reads a DNA methylation (or gene expression) data file in binary format |
| --befile-flist | reads a file to get the full paths of the binary files |
| --beta2m | calculates the methylation m value from the methylation beta value |
| --blup-probe | calculates the BLUP solutions for the probe effects |
| --chr | specifies a chromosome to select probes |
| --covar | reads discrete covariates from a plain text file |
| --detection-pval-file | reads a file that contains DNA methylation detection p-values |
| --dpval-mth | specifies a method to do quality control with the detection p-values |
| --dpval-thresh | specifies a threshold of detection p-value |
| --efile | reads a DNA methylation (or gene expression) data file in plain text format |
| --exclude-probe | excludes a subset of probes |
| --extract-probe | extracts a subset of probes |
| --from-probe | specifies the start probe |
| --from-probe-kb | specifies the start physical position of the probes |
| --gene | extracts a gene |
| --genes | extracts a subset of genes |
| --gene-expression | indicates gene expression profiles in the file |
| --get-mean | calculates the mean of each probe |
| --get-variance | calculates the variance of each probe |
| --keep | extracts a subset of individuals |
| --linear | saves linear regression statistics to a plain text file |
| --lower-beta | removes the DNA methylation probes with the mean beta value smaller than a specified threshold |
| --m2beta | calculates the methylation beta value from the methylation m value |
| --make-bod | saves DNA methylation (or gene expression) data in binary format |
| --make-efile | saves the DNA methylation (or gene expression) data in text format |
| --make-orm | estimates the omics relationship matrix (ORM) and save the lower triangle elements of ORM to binary files |
| --make-orm-bin | estimates the omics relationship matrix (ORM) and save the lower triangle elements of ORM to binary files |
| --make-orm-gz | estimates the omics relationship matrix (ORM) and save the lower triangle elements of ORM to compressed plain text files |
| --make-tefile | saves the DNA methylation (or gene expression) data in transposed text format |
| --multi-orm | reads multiple ORMs in binary format |
| --methylation-m | indicates methylation m values in the file |
| --methylation-beta | indicates methylation beta values in the file |
| --mlma | initiates an MLM based association analysis including the target probe (the probe to be tested for association) in the ORM |
| --mlma-exact | initiates an exact approach of MLM based association analysis with the chromosome where the target probe is located excluded from the ORM |
| --mpheno | reads a list of comma-delimited trait numbers if the phenotype file contains more than one trait |
| --missing-ratio-probe | specifies a missing proportion threshold to remove probes |
| --no-fid | indicates data without family ID |
| --orm | reads the ORM binary files |
| --orm-alg | specifies the algorithm to estimate the ORM |
| --orm-cutoff | removes one of a pair of individuals with estimated omics relationships larger than the specified cut-off value |
| --out | saves data (or results) in a file |
| --pca | conducts principal component analysis and saves the first n (default as 20) PCs |
| --pheno | reads phenotype data from a plain text file |
| --probe | extracts a specified probe |
| --probe-wind | defines a window centred on a specified probe |
| --probe-rm | excludes a specified probe |
| --qcovar | reads quantitative covariates from a plain text file |
| --ratio-probe | specifies a proportion threshold to remove probes |
| --ratio-sample | specifies a proportion threshold to remove individuals |
| --reml | performs REML (restricted maximum likelihood) analysis |
| --reml-alg | specifies the algorithm to do REML iterations |
| --reml-est-fix | displays the estimates of fixed effects on the screen |
| --reml-maxit | specifies the maximum number of iterations |
| --reml-no-lrt | turns off the LRT |
| --reml-pred-rand | predicts the random effects by the BLUP (best linear unbiased prediction) method |
| --remove | excludes a subset of individuals |
| --score | reads score files for probes and generates predicted omics profiles for individuals |
| --score-has-header | indicates probe score file has headers |
| --simu-causal-loci | reads a list of probes as causal probes |
| --simu-cc | simulates a case-control trait and specifies the number of cases and the number of controls |
| --simu-eff-mod | specifies whether or not to standardize the causal probe |
| --simu-rsq | specifies the proportion of variance in phenotype explained by the causal probes |
| --simu-k | specifies the disease prevalence. The default value is 0.1 if this option is not specified |
| --simu-qt | simulates a quantitative trait |
| --std | removes the probes with the standard deviation smaller than a specified threshold |
| --to-probe | specifies the end probe |
| --to-probe-kb | specifies the end physical position of the probes |
| --update-opi | reads a fully annotated .opi file |
| --upper-beta | removes the DNA methylation probes with the mean beta value larger than a specified threshold |
Data Resource
vQTL summary data
# vQTL summary data for 13 UKB traits
Wang et al.performed a genome-wide variance quantitative trait locus (vQTL) analysis of ~5.6 million variants on 348,501 unrelated individuals of European ancestry for 13 quantitative traits in the UK Biobank using median-based Levene’s method implemented by OSCA.
vQTL summary data for 13 UKB traits inCOJO format:13-QT-vQTL-summary-data.zip(1.4 GB)