Gene-ENviroment Interaction Estimator
The following packages are required on a Linux machine to compile and use the software package.
g++
cmake
make
git clone https://github.com/sriramlab/GENIE.git
cd GENIE
mkdir build
cd build/
cmake ..
make
An executable file named GENIE will be in the build folder after the installation steps. Run GENIE as follows:
./GENIE <command_line arguments>
If there are many annotations, then we recommend to run the following :
./GENIE_mem <command_line arguments>
GENIE_mem is slower than GENIE, but it uses less memory than GENIE.
GENIE_multi_pheno allows for the input of multi-trait phenotype file:
./GENIE_multi_pheno <command_line arguments>
Alternatively, you may run either GENIE, GENIE_mem, or GENIE_multi_pheno with a newline-separated config file:
./GENIE --config <config file>
When using a config file, the full keys (e.g., genotype="") must be used instead of shortcut flags (e.g., -g).
-h, --help Show this message and exit
-g, --genotype The path of PLINK BED genotype file
-p, --phenotype The path of phenotype file
-c, --covariate The path of covariate file
-a, --annot The path of input annotation for partitioned heritability
-o, --output Output file path (prefix)
-e, --environment The path of environment file
-m, --model Specification of the model. Currently there are 3 options:
1. additive genetic (G) effects only (arg: G)
The model reduces to RHE-mc (Pazokitoroudi et al. Nat Commun (2020). https://doi.org/10.1038/s41467-020-17576-9).
2. additive genetic (G) and gene-environment (GxE) effects (arg: G+GxE)
The model treats noise/environment effects as homogeneous.
3. additive genetic (G), gene-environment (GxE) and heterogeneous noise (NxE) effects (arg: G+GxE+NxE)
The model treats noise/environment effects as heterogeneous.
-k, --num-vec The number of random vectors (10 is recommended).
-jn, --num-jack The number of jackknife blocks (100 is recommended).
-t, --nthreads The number of threads for multithreading
-tr, --trace Flag for printing trace summary files (.tr) and metadata (.MN), compatible with SUM-RHE.
If this flag is set, a phenotype file is no longer required (a dummy phenotype without covariates will be used). Currently supports G-model only.
-v, --verbose Verbose mode; Output extra information (Normal equation, number of samples, etc.).
-eXa, --eXannot By default, GENIE fits a single GxE variance component. To partition the GxE component w.r.t the annotation file, add '-eXannot' flag.
-np, --norm-proj-pheno By default, the phenotype vector is standardized after regressing covariates. Turn this off by setting '--norm-proj-pheno 0'.
-i, --cov-add-intercept By default, a vector of ones is appended to the covariates (the intercept term). Turn this off by setting '--cov-add-intercept 0'.
Genotype: must be in PLINK BED format.
Phenotype: must have a header in the following format
(multiple phenotypes only supported by GENIE_multi_pheno):
FID IID name_of_pheno_1 name_of_pheno_2 . . . name_of_pheno_n
Covariate: must have a header in the following format:
FID IID name_of_cov_1 name_of_cov_2 . . . name_of_cov_n
Environment: must have a header in the following format:
FID IID name_of_env_1 name_of_env_2 . . . name_of_env_n
Annotation: must have M rows (M=number of SNPs) and K columns (K=number of annotations).
If SNP i belongs to annotation j, then there is "1" in row i and column j.
Otherwise, there is "0". (delimiter is " ")
1) The number and order of individuals must be the same in phenotype,
genotype, environment, and covariate files.
2) The number and order of SNPs must be the same in bim file and annotation file.
3) The annotation file does not have a header.
4) SNPs with MAF=0 must be excluded from the genotype file.
5) GENIE excludes individuals with NA or -9 values in the phenotype or
environment files from the analysis.
GENIE is able to fit single/multiple additive and GxE variance components.
1) SNPs are partitioned with respect to the annotation file.
2) By default, GENIE fits a single GxE variance component.
To partition the GxE component w.r.t the annotation file, please add "-eXannot" flag.
3) GENIE fits noise by environment (NxE) variance component (heterogeneous noise).
4) By default, GENIE adds environmental variables to covariates as fixed effects.
Sample files are provided in the example directory to ensure that everything works well. Look at test.sh file and run it :
chmod +x test.sh
./test.sh
To simulate phenotypes with GxE effects, see https://github.com/sriramlab/Simulator.
1. Ali Pazokitoroudi, Zhengtong Liu, Andrew Dahl, Noah Zaitlen, Saharon Rosset, Sriram Sankararaman.
AJHG (2024); doi: [10.1016/j.ajhg.2024.05.015](https://doi.org/10.1016/j.ajhg.2024.05.015)
2. Ali Pazokitoroudi, Yue Wu, Kathryn S. Burch, Kangcheng Hou, Aaron Zhou, Bogdan Pasaniuc, Sriram Sankararaman.
Nature Communications (2020); doi: [10.1038/s41467-020-17576-9](https://doi.org/10.1038/s41467-020-17576-9)
v1.1.1