Hiring in life sciences? Share your open positions with our professional community. Read more Close

Advertisement

Generating machine-learning-ready enzyme sequence-function data from microdroplet screening.

Created on 17 Nov 2026

Authors

Kieran Didi, Simon V Mathis, Matthew Penner, Michal Lihan, Michael M Bronstein, Pietro Lio, Florian Hollfelder, Maximilian Gantz

Published in

Methods in enzymology. Volume 737. Pages 347-378. Epub Oct 31, 2026.

Abstract

If enzyme engineering is to achieve the goal of making enzyme reactivity truly 'programmable', large, quantitative sequence-function datasets will be indispensable to train and validate predictive models. Droplet microfluidics can measure millions of genotype-bearing compartments, but quantitative sequence-function data with an accurate error estimate requires coordination between library construction, sorting, sequencing and analysis. This chapter describes a practical workflow for connecting full-length enzyme sequences to droplet phenotypes and converting the resulting barcode-assigned read counts into quantitative fitness estimates with defined uncertainty - turning droplet microfluidics into a data generation device. The approach is not restricted to conventional deep mutational scanning: beyond single mutations, libraries may contain combinatorially mutagenised variants of a parent enzyme, fully synthetic genes or sequences recovered from metagenomic libraries. A previous long-read deep mutational scanning (lrDMS) campaign followed by machine-learning that was used to generate combinatorial sequence-function data is used to illustrate a practical implementation of data harvesting from a library screening experiment. We give recommendations for experimental planning and execution covering oversampling, gating, library skew, library cloning and barcode design and emphasise experimental and computational quality controls that identify noise or bias in the dataset that would jeopardise interpretability when they propagate into a machine-learning dataset.

PMID:
43039627
Bibliographic data and abstract were imported from PubMed on 17 Nov 2026.

Read full publication at:
Please sign in to see all details.

Advertisement

Stats

  • Community rating n/a 0 votes
  • Reviewers' rating n/a 0 votes
  • Your rating

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

  • Recommendations n/a n/a positive of 0 vote(s)
  • Views 0
  • Comments 0

Recommended by

  • No recommendations yet.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.

Advertisement