SPEACH_AF

SPEACH_AF enables AlphaFold2 to sample alternate protein conformations by applying in silico mutagenesis to multiple sequence alignments (MSAs) to encode and reveal structural heterogeneity.

Key Features:

• MSA perturbation: Implements in silico mutagenesis of the input multiple sequence alignment to introduce sequence variations that encode potential conformational states.
• AlphaFold2 integration: Uses perturbed MSAs as inputs to AlphaFold2 to prompt sampling of alternative predicted structures beyond a single conformation.
• Conformational ensemble generation: Produces multiple predicted conformations per protein sequence to create structural ensembles for downstream analysis.
• Systematic modeling pipeline: Implements a reproducible modeling pipeline for generating and evaluating alternative conformations.
• Benchmarking: Benchmarks predictions against well-known examples of conformationally flexible proteins to assess reliability and accuracy.
• Membrane protein application: Applies the MSA perturbation and AlphaFold2 workflow specifically to explore conformational landscapes of membrane proteins.

Scientific Applications:

• Conformational landscape exploration: Maps alternative structural states of proteins, with emphasis on membrane protein dynamics.
• Structure-based drug design: Provides diverse structural ensembles to inform ligand docking and design strategies.
• Biochemical and biophysical studies: Supplies multiple predicted conformations for planning and interpreting biochemical assays and biophysical experiments.
• Protein dynamics and interaction analysis: Aids investigation of protein dynamics, function, and interaction mechanisms through ensemble-based structural hypotheses.

Methodology:

In silico mutagenesis is applied to the input MSA to generate perturbed alignments, which are then used as inputs to AlphaFold2 to produce multiple structure predictions; the workflow is organized as a systematic modeling pipeline and benchmarked against known flexible proteins.