Back

StructureDistiller: Structural relevance scoring increases resilience of contact maps to false positive predictions

Bittrich, S.; Schroeder, M.; Labudde, D.

2019-07-11 bioinformatics
10.1101/697839 bioRxiv
Show abstract

Protein folding and structure prediction are two sides of the same coin. We propose contact maps and the related techniques of constraint-based structure reconstruction as unifying aspect of both processes. The presented Structural Relevance (SR) score quantifies the contribution of individual contacts and residues to structural integrity.\n\nIt is demonstrated that entries of a contact map are not equally relevant for structural integrity. Structure prediction methods should explicitly consider the most relevant contacts for optimal performance because they effectively double resilience toward false positively predicted contacts. Furthermore, knowledge of the most relevant contacts significantly increases reconstruction fidelity on sparse contact maps by 0.4 [A].\n\nProtein folding is commonly characterized with spatial and temporal resolution: some residues are Early Folding while others are Highly Stable with respect to unfolding events. Using the proposed SR score, we demonstrate that folding initiation and structure stabilization are distinct processes.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.