Adversis: Adverse Response Signature Inference System.

Landing Page
Find Effects Page

Inspiration

Large scale transcriptome classification and interactions to better understand the metabolism of the human body and how it interacts with exogenous inputs.

What it does

Uses a model trained on scRNA - drug interaction data to predict potential interactions between drug agents and genes.

How we built it

Using the TAHOE-100M dataset as a base to refine modelling on drug-gene interactions

Challenges we ran into

Data availability, computational load limitations, and computational speed.

Accomplishments that we're proud of

We managed to make a trained model that fed into a front page website with searchable information

What we learned

New systems for large scale data analysis (AWS).

What's next for Adversis: Adverse Response Signature Inference System.

refining the model with deep learning, run model with multiple drugs in sequence, expand to bulk cell interactions as well as single cell, and train/test with context-aware setup (known side effects, pathway activities, cell lineages, dosage levels).