Discovery of drug–omics associations in type 2 diabetes with generative deep-learning models

Abstract

The application of multiple omics technologies in biomedical cohorts has the potential to reveal patient-level disease characteristics and individualized response to treatment. However, the scale and heterogeneous nature of multi-modal data makes integration and inference a non-trivial task. We developed a deep-learning-based framework, multi-omics variational autoencoders (MOVE), to integrate such data and applied it to a cohort of 789 people with newly diagnosed type 2 diabetes with deep multi-omics phenotyping from the DIRECT consortium. Using in silico perturbations, we identified drug–omics associations across the multi-modal datasets for the 20 most prevalent drugs given to people with type 2 diabetes with substantially higher sensitivity than univariate statistical tests. From these, we among others, identified novel associations between metformin and the gut microbiota as well as opposite molecular responses for the two statins, simvastatin and atorvastatin. We used the associations to quantify drug–drug similarities, assess the degree of polypharmacy and conclude that drug effects are distributed across the multi-omics modalities.

A machine learning and deep learning-based integrated multi-omics technique for leukemia prediction

“…The relationship among variables can be measured with values that fall from −1 to 1. To assists the understanding of correlation among variable, the value or coefficient is scaled within this range [ 67 ]. If the value of the coefficient is 0, that implies there is no link between variables A and B. Conversely, if the coefficient is between −1 and 1, either A to B or B to A is precisely predicted (see Fig.…”

Nanomedomics

“…These findings can provide an in-depth understanding of biological processes and therefore foster the development of improved therapeutic agents (Figure ). Capitalizing on this, multiomics analyses have become essential tools for molecular compounds. − Hundreds of molecular targeted drugs have been developed and approved by the FDA over the past decade, highlighting the enormous impact of multiomics on life sciences.…”

Adopting Mechanistic Molecular Biology Approaches in Exposome Research for Causal Understanding

“…Multiomics data sets from longitudinal human cohorts are beginning to provide new opportunities to establish more substantial and grounded evidence of biomolecular responses to exposures, ,,− but conducting omics profiling (metabolomics, genomics, transcriptomics, proteomics, microbiome) is limited by financial constraints. Predicting the total cost of complete biomolecular profiling per sample and the feasibility of omics at scale is dependent on multiple factors, including whether samples are analyzed in industry vs academic settings as well as the availability of commercial kits or providers.…”