Geometric Deep Learning Framework for De Novo Genome Assembly

Explore a groundbreaking approach to de novo genome assembly in this hour-long conference talk. Delve into GNNome, a geometric deep learning framework that revolutionizes path identification in assembly graphs. Learn how this innovative method overcomes challenges posed by repetitive regions, resulting in more contiguous assemblies across various species. Discover the framework's ability to leverage symmetries inherent to the problem and its performance compared to state-of-the-art tools. Gain insights into the potential of GNNome as a cornerstone for future work on complex genome reconstruction, including those with different ploidy and aneuploidy degrees. Understand the advantages of combining simulated data generation with AI approaches in genomic research. Access information on the publicly available framework and best-performing model for assembling new haploid genomes. Connect with the speaker, Lovro Vrček, a PhD student at the University of Zagreb, and join the AI for drug discovery community through the provided Portal link.

Geometric deep learning framework for de novo genome assembly | Lovro Vrček