Dataset Information

Navigating high-order protein mutational landscapes via deep learning on directed evolution trajectories

ABSTRACT: Accurately predicting the fitness effects of high-order mutations is a grand challenge in understanding and engineering proteins. Existing models, including pre-trained protein language models, struggle to capture the multi-residue interactions that govern these effects. Here, we introduce DENet, a deep learning framework that harnesses the rich co-mutation information within directed evolution (DE) trajectories to reconstruct high-resolution fitness landscapes for deciphering and engineering of complex protein variants. Applied to the cancer target KRAS, DENet-guided screening systematically identified high-order mutants with potent activities and uncovered hidden allosteric mechanisms. For MEK1, DENet discovered complex variants with >1,000-fold increased drug resistance, revealed synergistic tail mutations, and retrospectively identified over 75% of known clinical mutations, largely outperforming existing models. To broaden the framework’s applicability, we developed an in silico strategy that simulates directed evolution to generate crucial co-mutation information from widely available single-mutant datasets. DENet provides a quantitative framework for navigating complex fitness landscapes, uniting the rational engineering of multi-mutation proteins with the elucidation of their allosteric and clinical implications.

ORGANISM(S): Mesocricetus auratus

PROVIDER: GSE315318 | GEO | 2026/01/15

REPOSITORIES: GEO

ACCESS DATA

Dataset's files

Source:

			Action	DRS
		Other

Items per page:

1 - 1 of 1

Similar Datasets

Project description:Allosteric communication between distant sites in proteins is central to nearly all biological regulation but still poorly characterised for most proteins, limiting conceptual understanding, biological engineering and allosteric drug development. Typically only a few allosteric sites are known in model proteins, but theoretical, evolutionary and some experimental studies suggest they may be much more widely distributed. An important reason why allostery remains poorly characterised is the lack of methods to systematically quantify long-range communication in diverse proteins. Here we address this shortcoming by developing a method that uses deep mutational scanning to comprehensively map the allosteric landscapes of protein interaction domains. The key concept of the approach is the use of ‘multidimensional mutagenesis’: mutational effects are quantified for multiple molecular phenotypes—here binding and protein abundance—and in multiple genetic backgrounds. This is an efficient experimental design that allows the underlying causal biophysical effects of mutations to be accurately inferred en masse by fitting thermodynamic models using neural networks. We apply the approach to two of the most common human protein interaction domains, an SH3 domain and a PDZ domain, to produce the first global atlases of allosteric mutations for any proteins. Allosteric mutations are widely dispersed with extensive long-range tuning of binding affinity and a large mutational target space of network-altering ‘edgetic’ variants. Mutations are more likely to be allosteric closer to binding interfaces, at Glycines in secondary structure elements and at particular sites including a chain of residues connecting to an opposite surface in the PDZ domain. This general approach of quantifying mutational effects for multiple molecular phenotypes and in multiple genetic backgrounds should allow the energetic and allosteric landscapes of many proteins to be rapidly and comprehensively mapped.

Dataset Information

Navigating high-order protein mutational landscapes via deep learning on directed evolution trajectories

Dataset's files

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets