The flowchart highlights the key stages and methodologies of data generation and analyses across the three phases of FarmGTEx, including genomic variant calling (a), definition and quantification of molecular phenotypes (b), missing data imputation (c), batch effect inference and removal (d), and molQTL mapping via, for example, OmiGA (https://omiga.farmgtex.org) and ClipperQTL (e) 47, 68. Examples of approaches and methods used for each stage are shown in parentheses. ATAC-seq, assay for transposase-accessible chromatin with sequencing; dap-g, deterministic approximation of posteriors: genetics 69; DNABERT 70; DNAGPT 71; DNA-seq, DNA sequencing; GC–MS, gas chromatography–mass spectrometry; GCTA, a tool for genome-wide complex trait analysis 72; GRM, genetic relationship matrix; GLIMPSE, genotype likelihoods imputation and phasing method 73; LC–MS, liquid chromatography–mass spectrometry; PC, principal component; PLINK, a tool for whole-genome association and linkage analyses 74; phASER, phasing and allele-specific expression from RNA-seq 75; SuSiEx, an enhanced tool for cross-ancestry fine-mapping of causal variants based on sum of single effects 76; WGBS, whole-genome bisulfite sequencing; WGCNA, weighted gene coexpression network analysis.
