Cross-species analysis of enhancer logic using deep learning

Liesbeth Minnoye (1) , Ibrahim Ihsan Taskiran (1) , David Mauduit (1) , Maurizio Fazio (2, 3) , Linde van Aerschot (1) , Gert Hulselmans (1) , Valerie Christiaens (1) , Samira Makhzami (1) , Monika Seltenhammer (4) , Panagiotis Karras (1) , Aline Primot (5) , Edouard Cadieu (5) , Ellen Van Rooijen (2, 3) , Jean-Christophe Marine (1) , Giorgia Egidy (6) , Ghanem Elias Ghanem (7) , Leonard Zon (2, 3) , Jasper Wouters (1) , Stein Aerts (1)

1 KU Leuven - Catholic University of Leuven - Katholieke Universiteit Leuven
2 Dana-Farber Cancer Institute [Boston]
3 Boston Children's Hospital
4 Medizinische Universität Wien = Medical University of Vienna
5 IGDR - Institut de Génétique et Développement de Rennes
6 Université Paris-Saclay
7 ULB - Université libre de Bruxelles

Deciphering the genomic regulatory code of enhancers is a key challenge in biology because this code underlies cellular identity. A better understanding of how enhancers work will improve the interpretation of noncoding genome variation and empower the generation of cell type-specific drivers for gene therapy. Here, we explore the combination of deep learning and cross-species chromatin accessibility profiling to build explainable enhancer models. We apply this strategy to decipher the enhancer code in melanoma, a relevant case study owing to the presence of distinct melanoma cell states. We trained and validated a deep learning model, called DeepMEL, using chromatin accessibility data of 26 melanoma samples across six different species. We show the accuracy of DeepMEL predictions on the CAGI5 challenge, where it significantly outperforms existing models on the melanoma enhancer of IRF4. Next, we exploit DeepMEL to analyze enhancer architectures and identify accurate transcription factor binding sites for the core regulatory complexes in the two different melanoma states, with distinct roles for each transcription factor, in terms of nucleosome displacement or enhancer activation. Finally, DeepMEL identifies orthologous enhancers across distantly related species, where sequence alignment fails, and the model highlights specific nucleotide substitutions that underlie enhancer turnover. DeepMEL can be used from the Kipoi database to predict and optimize candidate enhancers and to prioritize enhancer mutations. In addition, our computational strategy can be applied to other cancer or normal cell types.

Domains: Life Sciences [q-bio]

Liesbeth Minnoye, Ibrahim Ihsan Taskiran, David Mauduit, Maurizio Fazio, Linde van Aerschot, et al.. Cross-species analysis of enhancer logic using deep learning. Genome Research, 2020, 30 (12), pp.1815-1834. ⟨10.1101/gr.260844.120⟩. ⟨hal-02961183⟩