Fig. 2: Martini3-IDP performance in MDP systems.

A Mean Rg calculated from Martini3-IDP simulations plotted against experimental Rg. Rg data are presented as mean values +/- statistical error estimates, error bars of experimental Rg values were determined by a fitting error from the Guinier fit and error bars of simulations represent the standard deviation of 5 blocks from blocking analysis. B Full-length hnRNPA1 protein mean Rg calculated from Martini3-IDP simulations under different salt concentrations. Rg data are presented as mean values ± statistical error estimates, error bars of simulations represent the standard deviation of 5 blocks from blocking analysis. Snapshots of the salt concentration dependent behavior of full length hnRNPA1 protein are shown. Increased electrostatic screening undermines the electrostatic attraction between RRM and LCD region and releases the LCD tail of hnRNPA1. Residues of the RRM region are colored as pink; the LCD region is colored as cyan. C Normalized contact frequency along HMGB1 sequence (averaged across three simulations). The integrated contact population of boxA residue 70–90 region equals 25.8% in tailless HMGB1, significantly larger than the 12.3% in full-length HMGB1, although the contact population of the residue 150–180 region in box B in tailless HMGB1 is also slightly increased to 17.4% compared to 12.7% in presence of acid tails. This data shows that in the absence of the HMGB1 acid tail, this hetero-complex CXCL12-HMGB1 has a stronger binding preference for the HMGB1 boxA ___domain. D Snapshots of full length (left) and tailless (right) HMGB1 and CXCL12 complex. BoxA in HMGB1 is colored pink, the boxB region is colored orange, and the acid tail is red; CXCL12 is colored cyan. The acid tail interferes with the binding between boxA and CXCL12.