The genomic DNA is transcribed into mRNA, which is then translated into protein. After transcription, RNA undergoes a series of processing steps. RNA splicing is one of them in which some sections of the pre-mRNA (introns) are removed with the joining of the adjacent region (exons). The same RNA transcript can be differentially spliced into alternative isoforms with potentially different functions. About 95% of multi-exon genes in humans undergo alternative splicing. Which features govern splice-site choice? At present, we know that consensus splice sites (GU/AG) 5’ and 3’ splice sites, a polypyrimidine tract, and a branch point are required for a splicing reaction. Nevertheless, when multiple splice sites are present, how exactly splicing decisions are made is largely unknown. In addition, genetic variation can modulate splice-site choice, although the specific link between genetic variation and splicing decision is still very difficult to infer. We use a unique approach to understand the genomic determinants of splicing that govern the splice site choice in humans. We have developed SpliSER (Splice-site strength/usage Estimate from RNA-Seq), which allows empirical quantification of the usage of individual splice-sites from RNA-seq data. This quantification can then be used as a phenotype to carry out SpliSER-GWAS to map genetic variation associated with differences in splice-site usage. We are applying SpliSER-GWAS on RNA-Seq data from Genotype-Tissue Expression (The GTEx Consortium, 2020). I will present GWAS results for heart atrial tissue and also demonstrate how SpliSER-GWAS performs much better than current cutting-edge tools in linking splicing variation and genetic variation. Genome-wide comparison of splicing patterns associated with genotypic variation in humans can help us unravel the general principles and mechanisms associated with splicing.