We Built NeuroAutomata: protein variant effect prediction (axonagentic.ai)

NeuroAutomata, a new AI tool for predicting the effects of protein variations, has been developed using Meta AI's ESM-2 model. This innovation stems from a need in clinical settings where outdated tools like PolyPhen-2 hinder timely decision-making due to their cumbersome setup requirements and limited accuracy. Unlike its predecessors, NeuroAutomata offers a user-friendly interface that allows users to paste protein sequences directly, yielding predictive scores in under 30 seconds without requiring extensive programming or GPU access. This is a significant step for the AI/ML community, as it democratizes access to advanced protein analysis capabilities, making it much easier for researchers and clinicians to assess variants of uncertain significance (VUS) quickly. The tool has demonstrated strong validation results, achieving correlation scores (rho) surpassing those of legacy models in tests against diverse protein datasets. By integrating AI verification methods that assess the reliability of claims, NeuroAutomata showcases how modern AI can enhance scientific workflows while emphasizing the importance of experimental validation, marking a progressive evolution in genetic variant analysis.