NLM DIR Seminar Schedule

Efficient predictions of alternative protein conformations by AlphaFold2-based sequence association

The many successes of AlphaFold2 (AF2) have inspired methods to predict multiple protein conformations, many of which have biological significance. These methods assume that AF2 uses coevolutionary information to predict alternative protein conformations, but they perform poorly on fold-switching proteins, which remodel their secondary structures and modulate their functions in response to cellular stimuli. Here, we present a method designed to leverage AF2’s learning of protein structure more than coevolutionary inference. This method–called CF-random–outperforms other methods for predicting alternative conformations of not only fold switchers but also dozens of other proteins that undergo rigid body motions and local conformational rearrangements. CF-random captures multiple conformations more frequently and requires 3-8x less sampling than all other methods. It also enabled predictions of fold-switched assemblies unpredicted by AlphaFold3. Several lines of evidence indicate that CF-random works by sequence association, suggesting that training-set structures and sequences play an important role in which conformations can be predicted readily. This observation inspired a blind prediction mode for alternative protein conformations. We release CF-random for community use, specifying its strengths and limitations.

GeneAgent: Self-verification Language Agent for Gene Set Analysis using Domain Databases

Gene set analysis allows researchers to explore groups of genes that likely act together in specific biological processes or molecular functions. Recent work in gene set analysis has shown promising performance utilizing large language models (LLMs). Nonetheless, their results are subject to limitations common in LLMs, such as hallucinations. In response, we develop GeneAgent, the first language agent for gene set analysis that self-verifies by autonomously interacting with biological databases, reducing hallucinations and enhancing accuracy. GeneAgent generates novel function names or aligns with notable enriched terms for input gene sets. Benchmarking on 1,106 gene sets from different sources, GeneAgent consistently outperforms vanilla GPT-4 by a significant margin. A detailed manual review confirms the effectiveness of the self-verification module in minimizing hallucinations and generating a more reliable explanatory analysis. We also apply GeneAgent to seven novel gene sets derived from mouse B2905 melanoma cell lines, with expert evaluations showing that GeneAgent offers novel insights into gene functions and expediting knowledge discovery.

TBD