AMaLa: Analysis of Directed Evolution Experiments via Annealed Mutational Approximated Landscape

Area of Science:

• Computational Biology
• Evolutionary Biology
• Bioinformatics

Background:

• Directed Evolution experiments generate valuable data for understanding biological systems.
• Accurate inference of fitness landscapes is crucial for interpreting Directed Evolution results.
• Existing methods often rely on limited data from specific rounds or equilibrium assumptions.

Purpose of the Study:

• To introduce Annealed Mutational approximated Landscape (AMaLa), a novel method for inferring fitness landscapes.
• To leverage the complete temporal data from Directed Evolution experiments for more robust inference.
• To develop a model that accurately describes evolutionary dynamics and predicts phenotypic measures.

Main Methods:

• AMaLa utilizes time-dependent statistical weights to model sequence trajectories.
• The model incorporates an energy term for selection and a generalized Jukes-Cantor model for mutations.
• It assumes statistical sampling independence between sequenced rounds.

Main Results:

• AMaLa accurately describes Directed Evolution dynamics.
• The method successfully infers fitness landscapes that reproduce phenotypic measures, outperforming current strategies.
• The inferred models can predict structural properties of wild-type sequences.

Conclusions:

• AMaLa offers a significant advancement in analyzing Directed Evolution data.
• The method provides a more comprehensive and accurate approach to fitness landscape inference.
• This approach enhances our ability to understand and predict evolutionary processes and molecular properties.