Introduction

AI Genomics CollecTive (AIGCT) is a platform for systematically evaluating ML/AI models of variant effects across the spectrum of genomics-based precision medicine.

It consists of a database of variant effect data organized into categories based on the source of the data. We use the term, task, to refer to a category. For each task we have the following data:

• Binary label for each variant

• VEP scores for each variant for a variety of different VEP’s

The tasks available are:

• ClinVar

• Cancer - Hotspot and MSK Passenger

• ASD - Autism Spectrum Disorder

• DDD - Deciphering Developmental Disorders

• CHD - Congential Heart Disease

The platform provides a python package that supports the following functions:

• Browse and retrieve data from the database

• Evaluate the performance of a user AI/ML VEP’s against labels and VEP’s stored in our database.

• Generate a high level summary report of the analysis results.

• Extract the results of a performance evaluation to files for further analysis.

• Render the results in publication ready plots and figures.