Biology subjects

Li, F.-Z.

Publications and source records attributed to Li, F.-Z..

6 recordsLinked to original sources

bioRxiv2024

Evaluation of Machine Learning-Assisted Directed Evolution Across Diverse Combinatorial Landscapes

Various machine learning-assisted directed evolution (MLDE) strategies have been shown to identify high-fitness protein variants more efficiently than typical wet-lab directed evolution approaches. However, limited understanding of the factors influencing MLDE performance across diverse proteins has hindered optimal strategy selection for wet-lab campaigns. To address this, we systematically analyzed multiple MLDE strategies, including active learning and focused training using six distinct zero-shot predictors, across 16 diverse protein fitness landscapes. By quantifying landscape navigability with six attributes, we found that MLDE offers a greater advantage on landscapes which are more challenging for directed evolution, especially when focused training is combined with active learning. Despite varying levels of advantage across landscapes, focused training with zero-shot predictors leveraging distinct evolutionary, structural, and stability knowledge sources consistently outperforms random sampling for both binding interactions and enzyme activities. Our findings provide practical guidelines for selecting MLDE strategies for protein engineering.

bioengineering↗

bioRxiv2024

LevSeq: Rapid Generation of Sequence-Function Data for Directed Evolution and Machine Learning

Sequence-function data provides valuable information about the protein functional landscape, but is rarely obtained during directed evolution campaigns. Here, we present Long-read every variant Sequencing (LevSeq), a pipeline that combines a dual barcoding strategy with nanopore sequencing to rapidly generate sequence-function data for entire protein-coding genes. LevSeq integrates into existing protein engineering workflows and comes with open-source software for data analysis and visualization. The pipeline facilitates data-driven protein engineering by consolidating sequence-function data to inform directed evolution and provide the requisite data for machine learning-guided protein engineering (MLPE). LevSeq enables quality control of mutagenesis libraries prior to screening, which reduces time and resource costs. Simulation studies demonstrate LevSeqs ability to accurately detect variants under various experimental conditions. Finally, we show LevSeqs utility in engineering protoglobins for new-to-nature chemistry. Widespread adoption of LevSeq and sharing of the data will enhance our understanding of protein sequence-function landscapes and empower data-driven directed evolution.

biochemistry↗

bioRxiv2024

Feature Reuse and Scaling: Understanding Transfer Learning with Protein Language Models

Large pretrained protein language models (PLMs) have improved protein property and structure prediction from sequences via transfer learning, in which weights and representations from PLMs are repurposed for downstream tasks. Although PLMs have shown great promise, currently there is little understanding of how the features learned by pretraining relate to and are useful for downstream tasks. We perform a systematic analysis of transfer learning using PLMs, conducting 370 experiments across a comprehensive suite of factors including different downstream tasks, architectures, model sizes, model depths, and pretraining time. We observe that while almost all down-stream tasks do benefit from pretrained models compared to naive sequence representations, for the majority of tasks performance does not scale with pretraining, and instead relies on low-level features learned early in pretraining. Our results point to a mismatch between current PLM pretraining paradigms and most applications of these models, indicating a need for better pretraining methods.

bioengineering↗

bioRxiv2023

DeCOIL: Optimization of Degenerate Codon Libraries for Machine Learning-Assisted Protein Engineering

With advances in machine learning (ML)-assisted protein engineering, models based on data, biophysics, and natural evolution are being used to propose informed libraries of protein variants to explore. Synthesizing these libraries for experimental screens is a major bottleneck, as the cost of obtaining large numbers of exact gene sequences is often prohibitive. Degenerate codon (DC) libraries are a cost-effective alternative for generating combinatorial mutagenesis libraries where mutations are targeted to a handful of amino acid sites. However, existing computational methods to optimize DC libraries to include desired protein variants are not well suited to design libraries for ML-assisted protein engineering. To address these drawbacks, we present DEgenerate Codon Optimization for Informed Libraries (DeCOIL), a generalized method which directly optimizes DC libraries to be useful for protein engineering: to sample protein variants that are likely to have both high fitness and high diversity in the sequence search space. Using computational simulations and wet-lab experiments, we demonstrate that DeCOIL is effective across two specific case studies, with potential to be applied to many other use cases. DeCOIL offers several advantages over existing methods, as it is direct, easy-to-use, generalizable, and scalable. With accompanying software (https://github.com/jsunn-y/DeCOIL), DeCOIL can be readily implemented to generate desired informed libraries. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=114 SRC="FIGDIR/small/540424v1_ufig1.gif" ALT="Figure 1"> View larger version (29K): org.highwire.dtl.DTLVardef@16ce726org.highwire.dtl.DTLVardef@bb8a2aorg.highwire.dtl.DTLVardef@6be56borg.highwire.dtl.DTLVardef@1a2efb6_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioengineering↗

bioRxiv2021

Two-dimensional multiplexed assay for rapid and deep SARS-CoV-2 serology profiling and for machine learning prediction of neutralization capacity

Antibody responses serve as the primary protection against SARS-CoV-2 infection through neutralization of viral entry into cells. We have developed a two-dimensional multiplex bead binding assay (2D-MBBA) that quantifies multiple antibody isotypes against multiple antigens from a single measurement. Here, we applied our assay to profile IgG, IgM and IgA levels against the spike antigen, its receptor-binding domain and natural and designed mutants. Machine learning algorithms trained on the 2D-MBBA data substantially improve the prediction of neutralization capacity against the authentic SARS-CoV-2 virus of serum samples of convalescent patients. The algorithms also helped identify a set of antibody isotype-antigen datasets that contributed to the prediction, which included those targeting regions outside the receptor-binding interface of the spike protein. We applied the assay to profile samples from vaccinated, immune-compromised patients, which revealed differences in the antibody profiles between convalescent and vaccinated samples. Our approach can rapidly provide deep antibody profiles and neutralization prediction from essentially a drop of blood without the need of BSL-3 access and provides insights into the nature of neutralizing antibodies. It may be further developed for evaluating neutralizing capacity for new variants and future pathogens.

immunology↗

bioRxiv2019

Chromatin 3D structure reconstruction with consideration of adjacency relationship among genomic loci

Chromatin 3D conformation plays important roles in regulating gene or protein functions. High-throughout chromosome conformation capture (3C)-based technologies, such as Hi-C, have been exploited to acquire the contact frequencies among genomic loci at genome-scale. Various computational tools have been proposed to recover the underlying chromatin 3D structures from in situ Hi-C contact map data. As connected residuals in a polymer, neighboring genomic loci have intrinsic mutual dependencies in building a 3D conformation. However, current methods seldom take this feature into account. We present a method called ShNeigh, which combines the classical MDS technique with local dependence of neighboring loci modelled by a Gaussian formula, to infer the best 3D structure from noisy and incomplete contact frequency matrices. The results obtained on simulations and real Hi-C data showed, while keeping the high-speed nature of classical MDS, ShNeigh is more accurate and robust than existing methods, especially for sparse contact maps. A Matlab implementation of the proposed method is available at https://github.com/fangzhen-li/ShNeigh.\n\nAuthor summaryWe propose a new method to infer a consensus 3D genome structure from a Hi-C contact map. The novelty of our method is that it takes into accounts the adjacency of genomic loci along chromosomes. Specifically, the proposed method penalizes the optimization problem of the classical multidimensional scaling method with a smoothness constraint weighted by a function of the genomic distance between the pairs of genomic loci. We demonstrate this optimization problem can still be solved efficiently by a classical multidimensional scaling method. We then show that the method can recover stable structures in high noise settings. We also show that it can reconstruct similar structures from data obtained using different restriction enzymes.

bioinformatics↗