Predicting Heteropolymer Phase Separation Using Two-Chain Contact Maps.

Area of Science:

• Biophysics and Soft Matter Physics
• Computational Biology and Biochemistry

Background:

• Phase separation in polymer solutions is often linked to single-chain ($R_{\text{g}}$) and two-chain ($B_{22}$) properties.
• Traditional metrics struggle to differentiate phase-separating from non-phase-separating heteropolymers, notably intrinsically disordered proteins (IDPs).

Purpose of the Study:

• To develop a predictive approach for heteropolymer phase separation using two-chain simulation data.
• To assess the efficacy of contact map analysis versus traditional metrics for predicting phase behavior.

Main Methods:

• Analysis of two-chain contact maps from simulations, quantifying monomer proximity.
• Training phase-separation classifiers using contact map statistics for minimal heteropolymer and residue-level IDP models.
• Comparison of contact map-based predictions against those derived from $R_{\text{g}}$ and $B_{22}$.

Main Results:

• Statistical properties of two-chain contact maps accurately predict heteropolymer phase separation.
• Contact map analysis significantly outperforms classifiers based solely on $R_{\text{g}}$ and $B_{22}$.
• Contact maps preserve crucial spatial interaction information absent in $B_{22}$.

Conclusions:

• Two-chain contact map analysis provides a robust and transferable method for predicting heteropolymer phase separation.
• This approach offers insights into the driving forces of IDP phase behavior based on physical interactions in dilute solutions.
• The developed method is computationally efficient for understanding complex biomolecular condensates.