Effective gene expression prediction from sequence by integrating long-range interactions

The next phase of genome biology research requires understanding how DNA sequence encodes phenotypes, from the molecular to organismal levels. How noncoding DNA determines gene expression in different cell types is a major unsolved problem, and critical downstream applications in human genetics depend on improved solutions. Here, we report substantially improved gene expression prediction accuracy from DNA sequence through the use of a new deep learning architecture called Enformer that is able to integrate long-range interactions (up to 100 kb away) in the genome. This improvement yielded more accurate variant effect predictions on gene expression for both natural genetic variants and saturation mutagenesis measured by massively parallel reporter assays. Notably, Enformer outperformed the best team on the critical assessment of genome interpretation (CAGI5) challenge for noncoding variant interpretation with no additional training. Furthermore, Enformer learned to predict promoter-enhancer interactions directly from DNA sequence competitively with methods that take direct experimental data as input. We expect that these advances will enable more effective fine-mapping of growing human disease associations to cell-type-specific gene regulatory mechanisms and provide a framework to interpret cis-regulatory evolution. To foster these downstream applications, we have made the pre-trained Enformer model openly available, and provide pre-computed effect predictions for all common variants in the 1000 Genomes dataset.One-sentence summaryImproved noncoding variant effect prediction and candidate enhancer prioritization from a more accurate sequence to expression model driven by extended long-range interaction modelling.

Gene-Wise Burden of Coding Variants Correlates to Noncoding Pharmacogenetic Risk Variants

Genetic variability can modulate individual drug responses. A significant portion of pharmacogenetic variants reside in the noncoding genome yet it is unclear if the noncoding variants directly influence protein function and expression or are present on a haplotype including a functionally relevant genetic variation (synthetic association). Gene-wise variant burden (GVB) is a gene-level measure of deleteriousness, reflecting the cumulative effects of deleterious coding variants, predicted in silico. To test potential associations between noncoding and coding pharmacogenetic variants, we computed a drug-level GVB for 5099 drugs from DrugBank for 2504 genomes of the 1000 Genomes Project and evaluated the correlation between the long-known noncoding variant-drug associations in PharmGKB, with functionally relevant rare and common coding variants aggregated into GVBs. We obtained the area under the receiver operating characteristics curve (AUC) by comparing the drug-level GVB ranks against the corresponding pharmacogenetic variants-drug associations in PharmGKB. We obtained high overall AUCs (0.710 ± 0.022–0.734 ± 0.018) for six different methods (i.e., SIFT, MutationTaster, Polyphen-2 HVAR, Polyphen-2 HDIV, phyloP, and GERP++), and further improved the ethnicity-specific validations (0.759 ± 0.066–0.791 ± 0.078). These results suggest that a significant portion of the long-known noncoding variant-drug associations can be explained as synthetic associations with rare and common coding variants burden of the corresponding pharmacogenes.