Introduction

popgen-cldla is a fully automated workflow to identify QTLs using combined linkage disequilibrium and linkage analysis (CLDLA). Unlike genome-wide association study (GWAS), cLDLA is a mixed-linear-model (MLM) based apporach to estimate variance component of each putative QTL using the specialized tools such as ASREML and blupf90+.

The workflow consists of the following steps:

1. Phase the genotypes using beagle or shapeit (optional)
2. Filtering the SNPs (optional)
3. Estimation of genetic relationship across genome (GRM)
4. Estimation of genetic relationship at each putative QTL region (DRM)
5. Bending and inverting GRM and DRM
6. Variance component estimation
7. Plotting the results in the form of manhattan plot

Note that the workflow currently only supports the singularity container

Input files

1. List of phased or unphased vcf files in a csv format

Bash 10,/data/testing/TailMLS04/OUT_VCF_BEAGLE4_ALL_OARAutoSom1_Chr10.TxT.vcf.gz,/data/testing/TailMLS04/OUT_VCF_BEAGLE4_ALL_OARAutoSom1_Chr10.TxT.vcf.gz.csi 11,/data/testing/TailMLS04/OUT_VCF_BEAGLE4_ALL_OARAutoSom1_Chr11.TxT.vcf.gz,/data/testing/TailMLS04/OUT_VCF_BEAGLE4_ALL_OARAutoSom1_Chr11.TxT.vcf.gz.csi . . .

1. Phenotype file

Bash iidx iid sex age bodywgt witherhgt taillgt 1 MLS169 1 11 10.6 47.0 32.7 2 MLS170 1 9 12.0 44.2 33.6 . . .

Note that the header should be present in the phenotype file and first, second and the last column in the phenotype file should be the index of individuals (with the header: iidx), individual id (with the header: iid) and the trait (Y), respectively.

1. Parameter file

Depending on the tool specified by the user, the workflow will generate the parameter file. things to consider in the parameter template file

1. If you have not used any of the two programs (asreml and blupf90+) implemented for variance component estimation before, please Refer to its respective documentation: asreml documentation or blupf90+ documentation. | 2. Note that in the current version, the MLMs are tested with two mixed effects : additive polygenic effect and additive QTL effects.

Example commands to run the workflow

To identify QTLs using CLDLA approach

Bash nextflow run cldla/ --input chrom_vcf_idx.csv --maf 0.05 --pheno_file TailMLS04.template.phe -qs 10 --outdir testing_blupf90_TailMLS04 -resume -profile singularity --output_prefix TailMLS04 --tool blupf90

To estimate heritability using the approach as implemented in GCTA

Bash nextflow run cldla/ --input nextflow_testing/TailMLS04/chrom_vcf_idx.csv --maf 0.05 --pheno_file nextflow_testing/TailMLS04/TailMLS04.template.phe -qs 10 --outdir testing_h2_TailMLS04 -resume -profile singularity --output_prefix TailMLS04_h2 --estimate_h2

Note that estimation of heritability using GCTA requires that the regressors be separated into two files: quantitative variables and qualitative variables. Therefore, in the phenotype file, any column with the float values (identified using the presence of dot,\".\") are automatically classified as quantitative and the column without float values are classified as qualitative. Further, the workflow to estimate h2, will produce the error if there is any column with mixtures of float and integer values.

Description of the parameters

Bash --input [string] Path to comma-separated file containing information about the samples in the experiment. --outdir [string] The output directory where the results will be saved. You have to use absolute paths to storage on Cloud infrastructure. --email [string] Email address for completion summary. --maf [number] minor allele frequency threshold; SNPs with MAF less than this threshold will filtered out --window_size [number] Window size to carry out cLDLA analysis [default: 40] --output_prefix [string] output prefix should not contain the dot in it [default: cldla_run1] --pheno_file [string] path to the phenotypes file as recognized by echidna --p_value [number] p-value cutoff based on permutation test --include_chrom [string] file containing ids of chromosome on which cLDLA will be carried out --num_autosomes [number] total number of autosomes in the dataset --estimate_h2 [boolean] whether to estimate heritability using gcta tool --lrt_threhold [number] lrt values above which the values are significant [default: 18] --tool [string] tool to estimate variance components [default: blupf90] --par_file [string] parameter file of the tool --n_perm [integer] number of permutation test to be carried out to determine the significant threshold for CLDLA or H2 estimation [default: 100] --phase_genotypes [boolean] whether or not to phase the genotypes --phasing_panel [string] csv file containing information about path to the vcf files to be used for imputation --phasing_map [string] csv file containing information about path to the recombination map files --phasing_tool [string] tool to be used for phasing: beagle5 or shapeit5 [default: beagle5]