Abstract 4879: Identification of degron motifs and mutations by analyzing a deep neural network

Abstract

Abstract Degrons, short amino acid sequences on protein substrates, play a critical role in the ubiquitin-proteasome system (UPS), which regulates over 80 percent of protein degradation in cells. When recognized by E3 ubiquitin ligases, degrons signal the destruction of its substrate protein, determining the regulatory specificity of the UPS. While a few degrons have been reported to be affected by mutations driving cancer growth, the effect of specific mutations on degrons remains largely unidentified. In this study, we analyze a deep neural network model that predicts the likelihood of an input protein sequence to contain a degron. To characterize the important C-terminal degron motifs learned by the model, we ran computational methods including sequence ranking and selection, enrichment analyses, and sequence logo visualization. This revealed the discovery of 101 significant degron motifs across lengths 2, 3, and 4. While previous studies suggested that di-amino acid motifs might be the primary source of c-end degron, the new motifs suggest that additional complexity might exist with a longer extended degron. Using these putative motifs as targets for mutations, we performed in silico mutagenesis to identify high-impact amino acid mutations that cause degron loss or gain. Our discovery of degron motif-mutation pairs could inform the development of targeted small-molecule drugs for cancer treatments and identification mutational biomarkers used for early detection of cancers. Citation Format: Cynthia Chen, Collin Tokheim, Shirley Liu. Identification of degron motifs and mutations by analyzing a deep neural network [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 4879.