Session at Pacific Symposium on Biocomputing Focuses on Finding Function in Biological Dark Matter

PNNL computational scientist Jason McDermott is kicking off the New Year with hundreds of like-minded individuals at the annual Pacific Symposium on Biocomputing (PSB). The symposium aims to capture databases, algorithms, interfaces, visualization, modeling, and other biocomputational methods from scientists around the world. McDermott is co-chairing one of only five organized sessions: Biological molecular function: methods and benchmarks for finding function in biological dark matter.

McDermott will be joined by session co-chairs Yana Bromberg (Emory University), Hannah Carter (University of California San Diego), and Travis Wheeler (University of Arizona).

“PSB is one of my favorite conferences: it’s a single track and offers many opportunities for interaction and networking,” said McDermott. “I’m excited this year to be bringing this topic to the community because I think it’s really timely and important.”

In the authors’ conference paper, they note that determining biological molecular function remains one of the most significant challenges in computational biology. One reason is the added mystery of dark matter in microbiomes, viruses, and unexplored sequence space, which complicates aspects like structure and function of proteins.

To illuminate the latest research and findings in these areas, the dark matter session includes an invited talk by Martin Steinegger (Seoul National University) and will highlight seven talks covering four scientific contributions:

• A geometric framework using signed distance functions for modeling protein surfaces.
• A reinforcement learning-based approach for steering protein generative models to design functional sequences.
• An ensemble framework combining sequence, structural, and network features for subcellular localization prediction.
• A scalable factorization method integrating gene-gene interaction data for analyzing high-dimensional genetic perturbation profiles.

At PNNL, McDermott leads a project that dovetails the PSB session topic. It’s part of the Predictive Phenomics Initiative (PPI), an internal investment that is focused on understanding the function of complex biological systems so researchers can engineer them for other uses. Dark matter permeates McDermott’s research into lipid production of oleaginous yeast, and he’s developing a computational framework for understanding the portions of molecules that are dark matter and how that affects their function.

“It’s an exciting time to be a computational scientist,” said McDermott, “Rapid advances in AI methods have transformed computational biology at the same time as there are leaps forward in methods for acquiring data about biological systems. The potential for new discoveries is higher than it’s ever been.”