Skip to content

Latest commit

 

History

History
594 lines (516 loc) · 24.3 KB

README.md

File metadata and controls

594 lines (516 loc) · 24.3 KB
Raw

G2P

A Genome-Wide-Association-Study Simulation Tool for Genotype Simulation, Phenotype Simulation, and Power Evaluation

More abundant simulation functions could be referred to our newly developed package SIMER for simulation of life science and breeding

Authors:

You Tang and Xiaolei Liu

Contact:

[[email protected]](Xiaolei Liu)

Contents

Installation

back to top

Environment Setup

back to top
JDK1.8 should be installed and environment variables must be configured before using G2P (http://www.oracle.com/technetwork/java/javase/downloads/jdk8-downloads-2133151.html)

Windows

back to top
GUI
Download all files from https://github.com/XiaoleiLiuBio/G2P/tree/master/gG2P_win_64 and double click the .jar file
Pipeline
Download all files from https://github.com/XiaoleiLiuBio/G2P/tree/master/kG2P_win_64

Mac

back to top
GUI
Download all files from https://github.com/XiaoleiLiuBio/G2P/tree/master/gG2P_mac and double click the .jar file
Pipeline
Download all files from https://github.com/XiaoleiLiuBio/G2P/tree/master/kG2P_mac
permission setting

$ chmod 777 gemma oldplink plink

Linux

back to top
GUI
Download all files from https://github.com/XiaoleiLiuBio/G2P/tree/master/gG2P_linux_x86_64 and run

$ Java -jar gG2P.jar

Pipeline
Download all files from https://github.com/XiaoleiLiuBio/G2P/tree/master/kG2P_linux_x86_64
permission setting

$ chmod 777 gemma oldplink plink

Data Preparation

All files should be prepared with the same prefix

ped

details see http://zzz.bwh.harvard.edu/plink/data.shtml#ped
back to top

Family ID Individual ID Father ID Mother ID Sex Trait marker 1 marker 2 marker 3 marker 4 marker 5 marker 6
1 33-16 0 0 0 2 0 0 A A A A A G A G A G
1 38-11 0 0 0 2 0 0 A G A G A A A G A G
1 4226 0 0 0 2 0 0 A G A A A A A G A G
1 4722 0 0 0 2 0 0 A G A G A A A G A G
1 A188 0 0 0 2 0 0 A A A A A A A G A G
1 A214N 0 0 0 2 0 0 A G A A A G A A A G
1 A239 0 0 0 2 0 0 A A A A A G A G A A
Family ID Individual ID Father ID Mother ID Sex Trait marker 1 marker 2 marker 3 marker 4 marker 5 marker 6
1 33-16 0 0 0 2 0 0 1 1 1 1 1 3 1 3 1 3
1 38-11 0 0 0 2 0 0 1 3 1 3 1 1 1 3 1 3
1 4226 0 0 0 2 0 0 1 3 1 1 1 1 1 3 1 3
1 4722 0 0 0 2 0 0 1 3 1 3 1 1 1 3 1 3
1 A188 0 0 0 2 0 0 1 1 1 1 1 1 1 3 1 3
1 A214N 0 0 0 2 0 0 1 3 1 1 1 3 1 1 1 3
1 A239 0 0 0 2 0 0 1 1 1 1 1 3 1 3 1 1

map

details see http://zzz.bwh.harvard.edu/plink/data.shtml#map
back to top

Chromosome ID Marker ID Genetic Distance Physical Distance
1 PZB00859.1 0 157104
1 PZA01271.1 0 1947984
1 PZA03613.2 0 2914066
1 PZA03613.1 0 2914171
1 PZA03614.2 0 2915078
1 PZA03614.1 0 2915242
1 PZA00258.3 0 2973508

pop

back to top
new samples will be generated using samples within sub-population

Sample ID sub-Population ID
33-16 1
38_11 1
4226 1
4722 2
A188 2
A214N 2
A239 2
A272 2
A441-5 2
A554 3
A556 3
A6 3
A619 3

qtn

back to top
each column represents simulated QTNs for each phenotype

Phenotype 1 Phenotype 2 Phenotype 3 Phenotype 4 Phenotype 5
66 67 80 83 90
9 15 52 59 135
90 96 143 147 174
3 3 15 58 89
89 118 185 203 212
69 72 72 84 110
46 59 125 204 207
14 15 19 29 39
9 23 65 111 131
19 52 74 179 194

Genotype Simulation

Single Population _ GUI

back to top

Ped: ped file
Map: map file
Path for output Ped/Map: path for output ped and map file
Block: Yes or No, if "Yes", the whole genome will be divided into blocks and shuffled to generate new samples
Number of SNPs in each block: Number of SNPs in each block
Mutation rate: the frequency of new mutations
Imputation: if TRUE, major allele will be used to impute missing values
Population size: simulated sample size

Single Population _ Pipeline

back to top

Windows

java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --outgen D:\data\output --rn 100 --block 4 –impute
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --outgen D:\data\output --rn 100 --block 4 --mutation 0.0001 --impute

Linux/Mac

java -jar kG2P.jar --ped /root/data/AG.ped --map /root/data/AG.map --outgen /root/data/output --rn 100 --block 4 –impute
java -jar kG2P.jar --ped /root/data/AG.ped --map /root/data/AG.map --outgen /root/data/output --rn 100 --mutation 0.0001
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --outgen D:\data\output --rn 100 --block 4
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --outgen D:\data\output --rn 100 --impute
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --outgen D:\data\output --rn 100
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --outgen D:\data\output --rn 100 --mutation 0.0001

jar: executive software
ped: ped file
map: map file
outgen: output path
block: number of SNPs in each block
rn: simulated sample size
impute: if 'impute' is added, major allele will be used to impute missing value
mutation: the frequency of new mutations

Multi Populations _ GUI

back to top

Ped: ped file
Map: map file
Pop: pop file
Path for output Ped/Map: path for output ped and map file
Block: Yes, or No, if "Yes", the whole genome will be divided into blocks and shuffled to generate new samples
Number of SNPs in each block: Number of SNPs in each block
Mutation rate: the frequency of new mutations
Migration rate: the ratio of immigrants (or emigrants) for each group
Genetic drift: is the change in the frequency of an existing gene variant (allele) in a population due to random sampling of organisms
Imputation: if TRUE, major allele will be used to impute missing values
Sample size of each population: sample size of each newly simulated population 
Population size: number or vector, simulated sample size

Multi Populations _ Pipeline

back to top

Windows

java -jar kG2P.jar --ped D:\data\AG.ped –map D:\data\AG.map --pop D:\data\AG.pop --outgen D:\data\output --block 4 --rn 100
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --pop D:\data\AG.pop --outgen D:\data\output --block 4 --rn 100 --mutation 0.0001 --mig 0.1 --genetic 0.001

Linux/Mac

java -jar kG2P.jar --ped /root/data/AG.ped --map /root/data/AG.map --pop /root/data/AG.pop --outgen /root/data/output --impute --block 4 --rn 100
java -jar kG2P.jar --ped /root/data/AG.ped --map /root/data/AG.map --pop /root/data/AG.pop --outgen /root/data/output --rn 100 --mutation 0.0001 --mig 0.1 --genetic 0.001
java -jar kG2P.jar --ped /root/data/AG.ped --map /root/data/AG.map --pop /root/data/AG.pop --outgen /root/data/output --rn 100 --genetic 0.001
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --pop D:\data\AG.pop --outgen D:\data\output --block 4 --rn 100
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --pop D:\data\AG.pop --outgen D:\data\output --rn 100
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --pop D:\data\AG.pop --outgen D:\data\output --impute --rn 100
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --pop D:\data\AG.pop --outgen D:\data\output --impute --block 4 --rn 100,200,300,400
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --pop D:\data\AG.pop --outgen D:\data\output --block 4 --rn 100,200,300,400
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --pop D:\data\AG.pop --outgen D:\data\output --rn 100,200,300,400
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --pop D:\data\AG.pop --outgen D:\data\output --impute --rn 100,200,300,400
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\AG.map --pop D:\data\AG.pop --outgen D:\data\output --rn 100 --mutation 0.0001 --mig 0.1 --genetic 0.001
java -jar kG2P.jar --ped D:\data\AG.ped --map D:\data\mydata.map --pop D:\data\mydata.pop --outgen D:\data\output –-rn 100 --genetic 0.003

jar: executive software
ped: ped file
map: map file
pop: pop file
outgen: output path
block: number of SNPs in each block
rn: simulated sample size
mutation: the frequency of new mutations
mig: the ratio of immigrants (or emigrants) for each group
genetic: genetic drift, which is the change in the frequency of an existing gene variant (allele) in a population due to random sampling of organisms
impute: if 'impute' is added, major allele will be used to impute missing values

Random Simulation _ GUI

back to top

Number of Chromosomes: total number of Chromosomes for each newly simulated sample
Marker size of each Chromosome: a vector, marker size of each Chromosome
Heterozygosity: average heterozygosity of all markers
Population size: Sample size of new generated population
Fst: fixation index
Physical distance of neighbor markers: Physical distance between neighbor markers
Output ped file path: output path of ped file

Random Simulation _ Pipeline

back to top

Windows

java -jar kG2P.jar --sample 100 --chr 5 --marker 100,200,300,400,500 --d 500 --outgen D:\data\output
java -jar kG2P.jar --sample 100,200,300 --chr 5 --marker 100,200,300,400,500 --d 500 --outgen D:\data\output --heter 0.5 --Fst 0.05,0.15,0.2

Linux/Mac

java -jar kG2P.jar --sample 100 --chr 5 --marker 100,200,300,400,500 --d 500 --outgen /root/data/output
java -jar kG2P.jar --sample 100,200,300 --chr 5 --marker 100,200,300,400,500 --d 500 --outgen /root/data/output --heter 0.5 --Fst 0.05,0.15,0.2

jar: executive software
sample: simulated sample size
chr: Number of Chromosomes
marker: SNP markers for each Chromosome
d: physical distance (base pairs) between nearby markers
outgen: output path
heter: average heterozygosity of all markers
Fst: fixation index

Phenotype Simulation

back to top

Phenotype _ GUI

Normal distribution

Geometry distribution

Ped file path: path of ped file
QTN area: the genomic area that used to select QTNs
Range: if 'QTN area' is 'Yes', 'range' can be used to set the 'QTN area'
Effect type: three options for QTN effect type, including A, A+D, A+D+A/D, A represents Additive effect and D represents Dominance effect
Distribution of QTN effects: two options, 'Normal' and 'Geometry'
Mean: mean value of the normal distribution, length of 'Mean' should be the same with 'Variance'
Variance: variance of the normal distribution
Formats: phenotype formats of 'GEMMA', 'Plink', and 'FaST-LMM' soft wares
Trait type: quantitative or category
Summary statistics: output the summary statistics information of simulated phenotypes, includes QTN ID, Chromosome, Physical Position, effects of additive, dominance, and additive-dominance interaction, and heritability 
Number of simulated phenotypes: number of simulated phenotypes
Number of QTNs: number of QTNs, if it is a vector, effect of different QTN group will follow different distributions; The length of 'Number of QTNs', 'Mean', and 'Variance' should be same
Heritability: heritability, a vector with length equals to number of ‘Effect type’
Output file path: output file path

Phenotype _ Pipeline

Windows

java -jar kG2P.jar --ped D:\data\AG.ped --outgen D:\data\output --rep 100 --dis geo 0.99 --h2 0.5 --nqtn 100 --QTNarea 1-500,1000-1500
java -jar kG2P.jar --ped D:\data\AG.ped --outgen D:\data\output --rep 10 --dis geo 0.99,0.88,0.77 --h2 0.3,0.1,0.05 --nqtn 80,40,20 --QTNarea 1-500,1000-1500 --effect A/D --category 2 --summary

Linux/Mac

java -jar kG2P.jar --ped /root/data/AG.ped --outgen /root/data/output --rep 100 --dis geo 0.99 --h2 0.5 --nqtn 100 --QTNarea 1-500,1000-1500
java -jar kG2P.jar --ped D:\data\AG.ped --outgen D:\data\Part2out --rep 100 --dis geo 0.99 --h2 0.5 --nqtn 100
java -jar .\kG2P.jar --ped D:\data\AG.ped --outgen D:\data\output --rep 10 --dis geo 0.99,0.88 --h2 0.5 --nqtn 10,20 --QTNarea 1-500,1000-1500
java -jar KG2P.jar --ped D:\data\AG.ped --outgen D:\data\Part2out --rep 100 --dis nor --m 0,0.1 --v 0.99,0.98 --h2 0.5 --nqtn 10,20 --QTNarea 1-500,1000-1500
java -jar kG2P.jar --ped /root/data/AG.ped --outgen /root/data/output --rep 10 --dis geo 0.99,0.88 --h2 0.3,0.1 --nqtn 80,40 --effect D --category 4
java -jar kG2P.jar --ped /root/data/AG.ped --outgen /root/data/output --rep 10 --dis nor --h2 0.5 --m 0 --v 0.99 --nqtn 100 --effect A --category 2
java -jar kG2P.jar --ped /root/data/AG.ped --outgen /root/data/output --rep 10 --dis nor --h2 0.3 --m 0.2 --v 0.88 --nqtn 100 –summary
java -jar kG2P.jar --ped /root/data/AG.ped --outgen /root/data/output --rep 10 --dis nor --h2 0.5 --m 0 --v 0.99 --nqtn 100 --effect A/D --category 2

jar: executive software
rep: number of simulated phenotypes
dis: distribution of QTN effects, two options, 'nor' and 'geo'
m: mean value of the normal distribution
v: variance of the normal distribution
QTNarea: the genomic area that used to select QTNs
h2: heritability
nqtn: number of QTNs, if it is a vector, the effects of different QTN groups will follow different distributions, and the length of nqtn, m, and v should be same
effect: three options for QTN effect type, including A (A), A+D (D), A+D+A/D (A/D)
category: number of categories for simulated phenotypes

Population Structure

back to top

Population structure _ GUI

Genotype (bed/bim/fam, ped/map): select genotype file
PCA: select if you want to calculate principle components
Number of PCs: number of PCs will be calculated and generated
Kinship: select if you want to calculate Kinship matrix

Population structure _ Pipeline

PC _ Windows

java -jar kG2P.jar --pre "plink --bfile D:\data\AG --pca 3 --out D:\data\AG"

PC _ Linux/Mac

java -jar kG2P.jar --pre "./plink --bfile /root/data/AG --pca 3 --out /root/data/AG"

Kinship _ Linux/Mac

java -jar kG2P.jar --pre "./gemma -bfile /root/data/AG -gk -o testgemma"

jar: executive software
pre: pipeline of the software you want to use, attention that the software should be put in the same path as kG2P.jar

Quality Control

back to top

Quality control _ GUI

Merge: merge multiple selected genotype file
Ped file path: path of ped file
Marker call rate: filters out all markers with missing call rates exceeding the provided value
Individual call rate: filters out all individuals with missing call rates exceeding the provided value
MAF: threshold of minor allele frequency
HWE: Hardy-Weinberg equilibrium test p-value threshold
Output file format: the format of output genotype file

Quality control _ Pipeline

Merge _ Windows

java -jar kG2P.jar --mergePED D:\data\mydata1.ped D:\data\mydata2.ped

Merge _ Linux/Mac

java -jar kG2P.jar --mergePED /root/data/mydata1.ped /root/data/mydat2.ped

QC _ Windows

java -jar .\kG2P.jar --ped "D:\data\mydata.ped" --geno 0.1 --mind 0.1 --maf 0.05 --hwe 0 --format pbv
java -jar .\kG2P.jar --ped "D:\data\mydata.ped" --geno 0.1 --mind 0.1 --maf 0.05 --hwe 0 --format b

QC _ Linux/Mac

java -jar ./kG2P.jar --ped /root/data/mydata.ped --geno 0.1 --mind 0.1 --maf 0.05 --hwe 0 --format pbv
java -jar kG2P.jar --ped /root/data/mydata.ped --geno 0.1 --mind 0.1 --maf 0.08 --hwe 0 --format bv

jar: executive software
mergePED: merge multiple genotype files prepared in ped format
geno: filters out all markers with missing call rates exceeding the provided value
mind: filters out all individuals with missing call rates exceeding the provided value
maf: threshold of minor allele frequency
hwe: Hardy-Weinberg equilibrium test p-value threshold
format: three options, including p(ped/map), b(bed/bim/fam), and v(vcf)

GWAS

back to top

GWAS _ GUI

Genotype (bed/bim/fam, ped/map): select genotype file
Phenotype file path: select the first phenotype file, all phenotypes in the same path will be analyzed one by one; name of the phenotype file must include a continuous order number, e.g., 'phenotype1.txt', 'phenotype2.txt', 'phenotype3.txt'
Results output file path: select output file path
Command: command for running gwas of the first phenotype, user-specific covariates files and kinship file can be also included in the command line, note that the output file name must be ‘g2ptemp’

R package _ Windows
Take FarmCPU (an R package) as an example:
Download and install essential packages for FarmCPU from http://www.zzlab.net/FarmCPU/index.html;
If the script for running FarmCPU is stored in Demo_script.R and the path of Rscript.exe is "E:\R\R-3.4.0\bin\x64\Rscript.exe", the command line should be written as: E:\R\R-3.4.0\bin\x64\Rscript.exe Demo_script.R

GWAS _ Pipeline

Plink _ Windows

java -jar kG2P.jar --GWAS "plink --bfile D:\data\AG --fam D:\data\out\171104010413\Plink\Plink_snps1.fam --assoc --out D:\data\g2ptemp" --sp Plink_snps1.fam

Plink _ Linux/Mac

java -jar kG2P.jar --GWAS "./plink --bfile /root/data/AG --fam /root/data/output/171104010413/Plink/Plink_snps1.fam --assoc --out /root/data/g2ptemp" --sp Plink_snps1.fam

Gemma _ Linux/Mac

java -jar kG2P.jar --GWAS "./gemma -bfile /root/data/AG -p /root/data/out/171104030401/GEMMA/GEMMA_phenotype1.txt -k /root/output/testgemma.cXX.txt -lmm 4 -o g2ptemp" --sp GEMMA_phenotype1.txt
jar: executive software
GWAS: command line used for running gwas of the first phenotype
sp: the first phenotype file, the file path is not needed

jar: executive software
GWAS: command line used for running gwas of the first phenotype
sp: the first phenotype file, the file path is not needed

Method Evaluation

back to top

Method Evaluation _ GUI

Map file: map file
QTN file: qtn file
GWAS result files path: path of gwas results
Column number of P values: column number of P values
Window size: A window is considered to be identified if a positive genetic marker is detected within the window
Output file path: output path of power/fdr evaluation results

Method Evaluation _ Pipeline

Evaluation of the methods in Plink _ Windows

java -jar kG2P.jar --map D:\data\AG.map --qtn D:\data\output\170106093742\qtn\test.qtn --gwas D:\data\output\Plink_snps1.qassoc --pv 9 --out D:\data\output
java -jar kG2P.jar --map D:\data\AG.map --qtn D:\oldoutput\190120151328\qtn\test.qtn --gwas D:\oldoutput\Plink_snps1.qassoc --pv 9 --out D:\data\output --WS 2,2000,200000,200000000

Evaluation of the methods in Plink _ Linux/Mac

java -jar kG2P.jar --map /root/data/AG.map --qtn /root/data/output/171104030401/qtn/test.qtn --gwas /root/data/output/Plink_snps1.qassoc --pv 9 --out /root/data/output
java -jar kG2P.jar --map /root/data/AG.map --qtn /root/data/output/171104030401/qtn/test.qtn --gwas /root/data/output/Plink_snps1.qassoc --pv 9 --out /root/data/output --WS 3,300,3000,30000,300000,3000000

Evaluation of the methods in Gemma _ Linux/Mac

java -jar kG2P.jar --map /root/data/AG.map --qtn /root/data/output/171104030401/qtn/test.qtn --gwas /root/output/GEMMA_phenotype1.assoc.txt --pv 9 --out /root/data/output

jar: executive software
map: map file
qtn: qtn file
gwas: the first gwas result file, there is a one-to-one mapping between gwas result files and columns in the qtn file
pv: column number of P values
out: output file path
WS: Window size, a window is considered to be identified if a positive genetic marker is detected within the window

Citations

For G2P: Please cite: You Tang, Xiaolei Liu; G2P: A Genome-Wide-Association-Study Simulation Tool for Genotype Simulation, Phenotype Simulation, and Power Evaluation, Bioinformatics, btz126, https://doi.org/10.1093/bioinformatics/btz126
For principle components analysis: 
    if you use plink, please also cite: Purcell S, et.al. "PLINK: A Tool Set for Whole-Genome Association and Population-Based Linkage  Analyses." American Journal of Human Genetics, 81.3(2007):559-575.
For calculating kinship matrix:
    if you use gemma, please also cite: Zhou, X., et.al. "Genome-wide Efficient Mixed Model Analysis for Association Studies." Nature Genetics, 44.7(2012):821.
Please cite all soft wares you used for GWAS and evaluation!

FAQ and Hints