Sampling Strategies for Experimentally Mapping Molecular Fitness Landscapes Using High-Throughput Methods.

Area of Science:

• Molecular Biology
• Evolutionary Biology
• Biophysics

Background:

• Mapping protein genotype-phenotype-fitness landscapes is crucial for understanding evolution.
• The vastness of genotype space makes comprehensive mapping challenging.
• Multiplexed assays of variant effect (MAVEs) enable high-throughput mapping but require optimized sampling strategies.

Purpose of the Study:

• To develop methods for approximating minimum sampling requirements for multi-mutational MAVEs.
• To experimentally validate these approximation methods.
• To compare sampling efficiency between nucleotide and amino acid variant simulations.

Main Methods:

• Developed computational methods and simulations to estimate sampling requirements for MAVEs.
• Created a novel library construction protocol to maximize sequence variation.
• Utilized ultradeep sequencing to validate library construction and experimental design.

Main Results:

• Demonstrated quantitative differences in sampling trajectories between nucleotide and amino acid variant simulations.
• Showed that nucleotide-based sampling requires greater effort for diminishing returns compared to amino acid-based sampling.
• Validated a new library construction protocol that efficiently captures extensive mutational combinations.

Conclusions:

• The proposed methods and library construction protocol aid in designing and upscaling pooled experimental screens.
• These advances facilitate broader exploration of protein genotype space.
• The findings are directly applicable to optimizing future MAVE experiments for evolutionary studies.