Bobola, N. | Search bioRxiv

bioRxiv2019

The epigenomic landscape regulating organogenesis in human embryos linked to developmental disorders

How the genome activates or silences transcriptional programmes governs organ formation. Little is known in human embryos undermining our ability to benchmark the fidelity of in vitro stem cell differentiation or cell programming, or interpret the pathogenicity of noncoding variation. Here, we studied histone modifications across thirteen tissues during human organogenesis. We integrated the data with transcription to build the first overview of how the human genome differentially regulates alternative organ fates including by repression. Promoters from nearly 20,000 genes partitioned into discrete states without showing bivalency. Key developmental gene sets were actively repressed outside of the appropriate organ. Candidate enhancers, functional in zebrafish, allowed imputation of tissue-specific and shared patterns of transcription factor binding. Overlaying more than 700 noncoding mutations from patients with developmental disorders allowed correlation to unanticipated target genes. Taken together, the data provide a new, comprehensive genomic framework for investigating normal and abnormal human development.

genomics↗

bioRxiv2019

Uncovering tissue-specific binding features from differential deep learning

MotivationTranscription factors (TFs) can bind DNA in a cooperative manner, enabling a mutual increase in occupancy. Through this type of interaction, alternative binding sites can be preferentially bound in different tissues to regulate tissue-specific expression programmes. Recently, deep learning models have become state-of-the-art in various pattern analysis tasks, including applications in the field of genomics. We therefore investigate the application of convolutional neural network (CNN) models to the discovery of sequence features determining cooperative and differential TF binding across tissues. ResultsWe analyse ChIP-seq data from MEIS, TFs which are broadly expressed across mouse branchial arches, and HOXA2, which is expressed in the second and more posterior branchial arches. By developing models predictive of MEIS differential binding in all three tissues we are able to accurately predict HOXA2 co-binding sites. We evaluate transfer-like and multitask approaches to regularising the high-dimensional classification task with a larger regression dataset, allowing for creation of deeper and more accurate models. We test the performance of perturbation and gradient-based attribution methods in identifying the HOXA2 sites from differential MEIS data. Our results show that deep regularised models significantly outperform shallow CNNs as well as k-mer methods in the discovery of tissue-specific sites bound in vivo. AvailabilityFor implementation and models please visit https://doi.org/10.5281/zenodo.2635463.

bioinformatics↗