evSeq: Cost-Effective Amplicon Sequencing of Every Variant in a Protein Library

Abstract

Widespread availability of protein sequence-fitness data would revolutionize both our biochemical understanding of proteins and our ability to engineer them. Unfortunately, even though thousands of protein variants are generated and evaluated for fitness during a typical protein engineering campaign, most are never sequenced, leaving a wealth of potential sequence-fitness information untapped. This largely stems from the fact that sequencing is unnecessary for many protein engineering strategies; the added cost and effort of sequencing is thus unjustified. Here, we present every variant sequencing (evSeq), an efficient protocol for sequencing a variable region within every variant gene produced during a protein engineering campaign at a cost of cents per variant. Execution of evSeq is simple, requires no sequencing experience to perform, relies only on resources and services typically available to biology labs, and slots neatly into existing protein engineering workflows. Analysis of evSeq data is likewise made simple by its accompanying software (found at github.com/fhalab/evSeq, documentation at fhalab.github.io/evSeq), which can be run on a personal laptop and was designed to be accessible to users with no computational experience. Low-cost and easy to use, evSeq makes collection of extensive protein variant sequence-fitness data practical.

Learning Epistasis and Residue Coevolution Patterns: Current Trends and Future Perspectives for Advancing Enzyme Engineering

“…3 Wittmann et al have recently published evSeq, a method for cost-efficient sequencing of variable regions within every variant of an engineering campaign based on computational tools and standardized components. 130 evSeq has the potential to drastically increase the amount of available sequence and fitness data, which would normally not be extracted during an enzyme engineering campaign, due to reasons of low time or cost efficiency. • Third, the different accuracies of predictions varying per model and assigned task point out more fundamental connections between data type or construction and inferred rules.…”

“…Models relying on experimental screening combined with machine learning might be better suited for such tasks. We will highlight a promising approach by Wittmann et al 130 in the following section.…”

Structure of Anabaena flos-aquae gas vesicles revealed by cryo-ET

“…Plasmid DNA was miniprepped (Econospin 96-well filter plate, Epoch Life Science) and verified by Sanger sequencing. Ultrasound-based phenotyping of mutants was performed in BL21-AI (Thermo) as previously described (Hurt et al, n.d.), and all screened mutants were sequenced using the evSeq pipeline (Wittmann et al 2022).…”

Inferring protein fitness landscapes from laboratory evolution experiments

“…Labeled protein data consist of a set of amino acid sequences and how each of those sequences map to a particular protein property of interest, such as thermostability, enzyme activity, or binding affinity. These sequence-function data are commonly generated using protein mutagenesis libraries and medium-or high-throughput assays to assign functional labels [5,6]. Supervised learning approaches such as linear regression or more sophisticated non-linear models can learn from labeled sequence-function data to infer the mapping from sequence to function [7][8][9][10].…”