A New Deep-Learning Model for Drug-Disease Interactions in T2D

Drug and Disease Interactions in Type II Diabetes

Complex diseases require complex treatments, and it is this common-sense approach that has given rise to the practice of “polypharmacy,” or the treatment of one condition with multiple drugs. It is challenging enough to characterize the variety of metabolomic, proteomic, and transcriptomic effects of one drug on a patient, let alone integrating the effects of multiple drugs.

In a recent Nature Biotechnology paper, Allesøe et. al. tackled the challenge of developing a deep learning model that could both integrate multi-omics patient data and learn connections between these -omics profiles and drugs used for treatment. They applied their research to a cohort of 789 newly diagnosed Type 2 diabetes (T2D) patients.

Integrating Multi-Omics Data

The first step was designing a deep-learning network based on variational autoencoders (VAE), which transform data from higher dimensions to a lower-dimensional space. During this process, the VAE can learn the structure of the input data and associations between the input variables, and thus has the ability to generalize or “fill in” missing data, a common challenge in integrating -omics and clinical datasets.

The Type 2 diabetes patient cohort was extensively characterized across genomics, transcriptomics, proteomics, metabolomics, and microbiomes as well as data on medication, diet questionnaires, and clinical measurements. The authors’ MOVE (multi-omics variational autoencoders) method successfully integrated multi-omics data with clinical data, resisted biases in the data and could account for missing data.

Extracting Drug and Multi-Omics Interaction Biomarkers

So what can a model like this tell us? The MOVE model identified an average of 20 -omics associations per drug, and was sensitive to all of the drugs given, whether they were commonly prescribed or more rare.

An example of the identified drug-data associations were between Type 2 diabetes biomarkers and metformin, one of the most commonly prescribed drugs for T2D. The authors identified 88 significant clinical/multi-omics interactions across their datasets. Specifically, 12 clinical biomaerkers of Type 2 diabetes including insulin clearance, activated GLP-1, and glucose sensitivity were significantly associated with metformin.

One of the goals of models like these is to drive precision medicine, or “N-of-1” treatments. With training on larger cohort sizes, researchers could leverage these models to determine the best treatments and outcomes for a patient based on their unique multi-omics profile upon presentation. Generative models like this can even go a step further and predict what happens when different inputs are perturbed.

Outsourcing Bioinformatics Analysis: How Bridge Informatics Can Help

Many of our clients at Bridge Informatics are pursuing these kinds of research questions with sophisticated bioinformatics approaches. From pipeline development and software engineering to deploying existing bioinformatics tools, Bridge Informatics can help you on every step of your research journey.As experts across data types from cutting-edge sequencing platforms, we can help you tackle the challenging computational tasks of storing, analyzing and interpreting genomic and transcriptomic data. Bridge Informatics’ bioinformaticians are trained bench biologists, so they understand the biological questions driving your computational analysis. Click here to schedule a free introductory call with a member of our team.

Jane Cook, Biochemist & Content Writer, Bridge Informatics

Jane Cook, leading Content Writer for Bridge Informatics, has written over 100 articles on the latest topics and trends for the bioinformatics community. Jane’s broad and deep interdisciplinary molecular biology experience spans developing biochemistry assays to genomics. Prior to joining Bridge, Jane held research assistant roles in biochemistry research labs across a variety of therapeutic areas. While obtaining her B.A. in Biochemistry from Trinity College in Dublin, Ireland, Jane also studied journalism at New York University’s Arthur L. Carter Journalism Institute. As a native Texan, she embraces any challenge that comes her way. Jane hails from Dallas but returns to Ireland any and every chance she gets. If you’re interested in reaching out, please email daniel.dacey@old.bridgeinformatics.com or dan.ryder@old.bridgeinformatics.com.

Recent Posts