Deep Learning in Pharmacogenomics: From Gene Regulation to Patient Stratification

Abstract

This Perspective provides examples of current and future applications of deep learning in pharmacogenomics, including: identification of novel regulatory variants located in noncoding domains of the genome and their function as applied to pharmacoepigenomics; patient stratification from medical records; and the mechanistic prediction of drug response, targets and their interactions. Deep learning encapsulates a family of machine learning algorithms that has transformed many important subfields of artificial intelligence over the last decade, and has demonstrated breakthrough performance improvements on a wide range of tasks in biomedicine. We anticipate that in the future, deep learning will be widely used to predict personalized drug response and optimize medication selection and dosing, using knowledge extracted from large and complex molecular, epidemiological, clinical and demographic datasets.

Hypothesis: Caco‐2 cell rotational 3D mechanogenomic turing patterns have clinical implications to colon crypts

“…We propose that modelling the 4D Nucleome dynamics, 4D mRNA distribution and actomyosin forces that regulate tight junction protein expression and function will predict the self‐organizing of epithelial cells in a cell type‐, developmental stage‐specific manner. This information will be useful in generating a precise mathematical model of human colon crypts, which could be employed as a powerful algorithm to help design precision medicine approaches for targeted, disease‐specific treatments in a variety of medical ailments, including functional bowel disorders (FBD) and colorectal cancer (CRC) …”

“…This information will be useful in generating a precise mathematical model of human colon crypts, which could be employed as a powerful algorithm to help design precision medicine approaches for targeted, disease-specific treatments in a variety of medical ailments, including functional bowel disorders (FBD) and colorectal cancer (CRC). 5,12,[14][15][16] To generate "proof of concept" data, we tracked the formation of a coordinated epithelial cell sheet during Caco-2 cell differentiation on a smooth, flat and hard glass surface that recapitulates known gene expression patterns that occur along the colon crypt axis. Detailed indepth description and discussion of the rotational 3D mechanogenomic Turing patterns observed during differentiation are included in the supplementary and online material ( Figure 1 and S1) (http:// www.socr.umich.edu/projects/3d-cell-morphometry/data.html).…”

Ketamine’s pharmacogenomic network in human brain contains sub-networks associated with glutamate neurotransmission and with neuroplasticity

“…This research emphasizes the role of the regulatory genome, including enhancer and superenhancer-based interactomes, as an approach that provides insight into pharmacogenomic network mechanisms (22)(23)(24)(25). It differs from pathway modeling methods in which altered protein folding based on missense codon variants and fixed signaling pathways serve as the foundation for the interpretation of the molecular substrate of ketamine response in humans.…”

“…The ketamine response workflow is based on the pharmacoepigenomics informatics pipeline (PIP) (22)(23)(24)(25). Input genes to the data analysis pipeline first included those genes that encode proteins and constituents of macromolecular protein complexes obtained from past studies of the binding affinity of (R, S)-ketamine, R-ketamine and S-ketamine performed in microsomal and tissue preparations in rodents and humans.…”

“…Our strategy analyzed disparate data from multiple scales of biological function combining bioinformatics, computational medicine and machine learning (22)(23)(24)(25). GWAS SNPs provided insight into the phenotypic consequences of the mutational alterations and variation within disease sub-networks impacted by ketamine, spatial interactions within chromatin that control expression of ketamine-response genes, and mesoscale functional neurocircuits within human brain regions that are rapidly activated following ketamine administration (Figure 1).…”