Deciphering regulatory DNA sequences and noncoding genetic variants using neural network models of massively parallel reporter assays

Abstract

The relationship between noncoding DNA sequence and gene expression is not well-understood. Massively parallel reporter assays (MPRAs), which quantify the regulatory activity of large libraries of DNA sequences in parallel, are a powerful approach to characterize this relationship. We present MPRA-DragoNN, a convolutional neural network (CNN)-based framework to predict and interpret the regulatory activity of DNA sequences as measured by MPRAs. While our method is generally applicable to a variety of MPRA designs, here we trained our model on the Sharpr-MPRA dataset that measures the activity of ∼500,000 constructs tiling 15,720 regulatory regions in human K562 and HepG2 cell lines. MPRA-DragoNN predictions were moderately correlated (Spearman ρ = 0.28) with measured activity and were within range of replicate concordance of the assay. State-of-the-art model interpretation methods revealed high-resolution predictive regulatory sequence features that overlapped transcription factor (TF) binding motifs. We used the model to investigate the cell type and chromatin state preferences of predictive TF motifs. We explored the ability of our model to predict the allelic effects of regulatory variants in an independent MPRA experiment and fine map putative functional SNPs in loci associated with lipid traits. Our results suggest that interpretable deep learning models trained on MPRA data have the potential to reveal meaningful patterns in regulatory DNA sequences and prioritize regulatory genetic variants, especially as larger, higher-quality datasets are produced.

Characterising the genetic basis of immune response variation to identify causal mechanisms underlying disease susceptibility

“…86 Massively parallel assays have also been used in the context of both splicing 88,89 and miRNA-mediated regulation, 90,91 showing, for instance, that up to 16% of splice disrupting variants are located in deep intronic regions. 88 With the development of deep learning frameworks for sequence-based predictions, 92,93 data generated by these assays, will now fuel the construction of predictive models. These models will, in turn, allow quantifying the regulatory impact of novel genetic variants on transcriptional activity, 86 alternative splicing, 94 or miRNA-mediated regulation.…”

Sequence-to-function deep learning frameworks for synthetic biology

“…Taking inspiration from work in the field of image recognition and genomics 26,[29][30][31] , we investigated the first convolutional layer to see which features our model deemed important by interpreting the filter weights learned from input sequences as sequence logos (Fig. 2E).…”

Functional characterization of thousands of type 2 diabetes-associated and chromatin-modulating variants under steady state and endoplasmic reticulum stress

“…Moreover, ATAC-seq peaks are characterized by 1 or more summits, which have higher chromatin accessibility than flanking genomic regions also within ATAC-seq peaks 6 . Few studies have explored the relationship between transcriptional activity and open chromatin regions defined by ATAC-seq 7, 8 . For instance, it is not known whether higher chromatin accessibility within ATAC-seq peaks also corresponds to enhanced transcriptional activity, or whether proximity to ATAC-seq peak summits influences the likelihood for a SNP to affect enhancer activity.…”