Resolving Interaction Specificity in Protein Paralogs

This direction is merged into another direction: direct coupling analysis. Its last version stays readable and citable here.

Version 1 of 116 papers20 conceptsfrozen 2026-10-04

In the world of structural biology, the "AlphaFold revolution" has largely mastered the art of predicting how a single protein chain folds. However, the next frontier is the 3D-proteome: understanding how thousands of different proteins in a cell recognize, bind, and assemble into functional complexes.

The primary tool for this task is sequence co-evolution. The intuition is simple: if two amino acids are in physical contact, they must evolve in lockstep. If one mutates, the other must also mutate to maintain the "handshake." By looking for these correlated mutations across thousands of species, researchers like Thomas A. Hopf and Chris Sander developed EVcomplex to predict the 3D structures of protein complexes from sequence alone.

But a massive technical hurdle remains: the paralog matching problem.

The Ambiguity of Protein Families

Most proteins do not exist in isolation; they belong to families of "paralogs"—closely related versions of the same protein within a single genome. For example, a human cell might contain dozens of nearly identical versions of a signaling protein.

When trying to map interactions between two protein families (Family A and Family B), the co-evolutionary signal becomes "noisy." If an organism has three versions of A (A1, A2, A3) and three versions of B (B1, B2, B3), which one interacts with which? As Anna G. Green and Oliver Kohlbacher have noted, incorrectly pairing these sequences during analysis leads to a total loss of the structural signal, making it impossible to see the physical interface.

Maximizing Signal through Iterative Matching

To solve this, researchers have turned the problem on its head. Instead of trying to guess the pairings first, they use the strength of the co-evolutionary signal itself to find the correct matches.

In Simultaneous identification of specifically interacting paralogs and interprotein contacts by direct coupling analysis, Carlo Baldassi, Martin Weigt, and Andrea Pagnani introduced algorithms that iteratively swap potential partners. The logic is that the "true" biological partners will exhibit the strongest co-evolutionary correlations. By maximizing this global signal, the algorithm simultaneously identifies which paralogs interact and which specific residues form the contact points. This allows for the de novo prediction of complex structures even when genomic proximity (which usually suggests interaction) is absent.

Biological Mechanisms of Specificity

While algorithms can untangle these signals, the proteins themselves have evolved clever physical mechanisms to ensure they only bind to the "right" partner.

The Lock and Key of Protocadherins

Neurons use a family of proteins called protocadherins as molecular barcodes to distinguish "self" from "non-self." With over 50 nearly identical isoforms, the risk of cross-reactivity is high. John M. Nicoludis and Rachelle Gaudet discovered that these proteins use a division of labor: certain regions provide general "glue" (affinity), while others act as a highly specific "lock and key" (specificity). This ensures that a neuron only repels a dendrite that expresses the exact same set of isoforms.

Dynamic Inhibition in SUMO

Other proteins use dynamic regulation to prevent accidental interactions. Sebastian M. Richter and colleagues found that the SUMO1 protein possesses a "floppy," intrinsically disordered N-terminal tail. This tail acts as an intramolecular inhibitor, transiently masking its own binding site. This "fuzzy" interaction provides a layer of selectivity, helping the cell distinguish between different SUMO variants that share nearly identical binding surfaces.

Toward the 3D Proteome

The ultimate goal of this research is to move from individual structures to a structurally resolved human interaction network. By combining high-accuracy folding models from AlphaFold2 with advanced co-evolutionary matching, researchers like Arne Elofsson and David F. Burke are beginning to map thousands of human interactions at residue-level resolution.

This shift allows us to see not just that two proteins interact, but how they interact, and how subtle mutations in a specific paralog might disrupt a signaling pathway or lead to disease. By solving the paralog matching problem, we are effectively clearing the "noise" from the evolutionary record to reveal the precise architecture of life's molecular machinery.

Papers in this version

History & citationVersion 1 · 2026-10-04

Merged into another direction · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_resolving_interaction_specificity_in_pro_f51f13c3,
  title = {Resolving Interaction Specificity in Protein Paralogs},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/resolving-interaction-specificity-in-protein-paralogs/rdr_f51f13c3df27430ab74c58f25fd58aac},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}

This exact version (version 1, frozen text):

@misc{lacuna_resolving_interaction_specificity_in_pro_f51f13c3_v1,
  title = {Resolving Interaction Specificity in Protein Paralogs},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/resolving-interaction-specificity-in-protein-paralogs/rdr_f51f13c3df27430ab74c58f25fd58aac/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-20},
}