Abstract:Elucidating how protein sequence determines the properties of disordered proteins and their phase-separated condensates is a great challenge in computational chemistry, biology, and biophysics. Quantitative molecular dynamics simulations and derived free energy values can in principle capture how a sequence encodes the chemical and biological properties of a protein. These calculations are, however, computationally demanding, even after reducing the representation by coarse-graining; exploring the large spaces of potentially relevant sequences remains a formidable task. We employ an "active learning" scheme introduced by Yang et al. (bioRxiv, 2022, https://doi.org/10.1101/2022.08.05.502972) to reduce the number of labelled examples needed from simulations, where a neural network-based model suggests the most useful examples for the next training cycle. Applying this Bayesian optimisation framework, we determine properties of protein sequences with coarse-grained molecular dynamics, which enables the network to establish sequence-property relationships for disordered proteins and their self-interactions and their interactions in phase-separated condensates. We show how iterative training with second virial coefficients derived from the simulations of disordered protein sequences leads to a rapid improvement in predicting peptide self-interactions. We employ this Bayesian approach to efficiently search for new sequences that bind to condensates of the disordered C-terminal domain (CTD) of RNA Polymerase II, by simulating molecular recognition of peptides to phase-separated condensates in coarse-grained molecular dynamics. By searching for protein sequences which prefer to self-interact rather than interact with another protein sequence we are able to shape the morphology of protein condensates and design multiphasic protein condensates.

Energy landscapes of homopolymeric RNAs revealed by deep unsupervised learning

The biogenesis of a cytochrome b complex in yeast mitochondria. Sites of translation of the protein components.

Sequence determinants of protein phase separation and recognition by protein phase-separated condensates through molecular dynamics and active learning

Exploring the energetic and conformational properties of the sequence space connecting naturally occurring RNA tetraloop receptor motifs

Nucleoprotein Phase-Separation Affinities Revealed via Atomistic Simulations of Short Peptide and RNA Fragments

Decoding Biomolecular Condensate Dynamics: An Energy Landscape Approach

RNA under Tension: Folding Landscapes, Kinetic Partitioning Mechanism, and Molecular Tensegrity

Free Energy Landscape and Multiple Folding Pathways of an H-Type RNA Pseudoknot

RNA secondary structure formation: a solvable model of heteropolymer folding

Phase Separation Modulates the Thermodynamics and Kinetics of RNA Hybridization

Statistical mechanics of secondary structures formed by random RNA sequences

Statistical-physics approaches to RNA molecules, families and networks

Phase-separating RNA-binding proteins form heterogeneous distributions of clusters in subsaturated solutions

Differential interactions determine anisotropies at interfaces of RNA-based biomolecular condensates

Accelerated Simulations Reveal Physicochemical Factors Governing Stability and Composition of RNA Clusters

Helix-Based RNA Landscape Partition and Alternative Secondary Structure Determination

RNA sequence and structure control assembly and function of RNA condensates

Sequence motif dynamics in RNA pools

Neolignan inhibitors of antigen-induced degranulation in RBL-2H3 cells from the needles of Pinus thunbergii.

Accelerated simulations of RNA clustering: a systematic study of repeat sequences