Learning protein fitness landscapes with deep mutational scanning data from multiple sources

Chen, Lin; Zhang, Zehong; Li, Zhenghao; Li, Rui; Huo, Ruifeng; Chen, Lifan; Wang, Dingyan; Luo, Xiaomin; Chen, Kaixian; Liao, Cangsong; Zheng, Mingyue

doi:10.1016/j.cels.2023.07.003

Cited by 4 publications

(6 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Regime extrapolation tests a model’s ability to predict how mutations combine by training on single amino acid substitutions and predicting the effects of multiple substitutions [18, 19, 21, 22] (Figs. 3c and S4).…”

Section: Resultsmentioning

confidence: 99%

Biophysics-based protein language models for protein engineering

Gelman,

Johnson,

Freschlin

et al. 2024

Preprint

View full text Add to dashboard Cite

Protein language models trained on evolutionary data have emerged as powerful tools for predictive problems involving protein sequence, structure, and function. However, these models overlook decades of research into biophysical factors governing protein function. We propose Mutational Effect Transfer Learning (METL), a protein language model framework that unites advanced machine learning and biophysical modeling. Using the METL framework, we pretrain transformer-based neural networks on biophysical simulation data to capture fundamental relationships between protein sequence, structure, and energetics. We finetune METL on experimental sequence-function data to harness these biophysical signals and apply them when predicting protein properties like thermostability, catalytic activity, and fluorescence. METL excels in challenging protein engineering tasks like generalizing from small training sets and position extrapolation, although existing methods that train on evolutionary signals remain powerful for many types of experimental assays. We demonstrate METL's ability to design functional green fluorescent protein variants when trained on only 64 examples, showcasing the potential of biophysics-based protein language models for protein engineering.

show abstract

Section: Resultsmentioning

confidence: 99%

Biophysics-based protein language models for protein engineering

Gelman,

Johnson,

Freschlin

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…types at an aggregate level (Strokach et al, 2021;Høie et al, 2022;Cagiada et al, 2023;Nguyen and Hy, 2023), although some results suggest that a richer representation might be learned by combining multiple data types at the input level (Mansoor et al, 2021;Wu et al, 2023;Wang et al, 2022;Yang et al, 2022;Chen et al, 2023;Cheng et al, 2023;Zhang et al, 2023).…”

Section: Gnnmentioning

confidence: 99%

“…Examples of the types of data used as input include the wild-type amino acid sequence ( Lin et al, 2022; Brandes et al, 2022 ), a multiple sequence alignment (MSA) ( Ng and Henikoff, 2001; Balakrishnan et al, 2011; Lui and Tiana, 2013; Nielsen et al, 2017; Hopf et al, 2017; Riesselman et al, 2018; Laine et al, 2019 ) or the protein structure ( Boomsma and Frellsen, 2017; Jing et al, 2021a; Hsu et al, 2022 ). Some methods have combined predictions from multiple protein data types at an aggregate level ( Strokach et al, 2021; Høie et al, 2022; Cagiada et al, 2023; Nguyen and Hy, 2023 ), although some results suggest that a richer representation might be learned by combining multiple data types at the input level ( Mansoor et al, 2021; Wu et al, 2023; Wang et al, 2022; Yang et al, 2022; Chen et al, 2023; Cheng et al, 2023; Zhang et al, 2023 ).…”

Section: Introductionmentioning

confidence: 99%

“…By constraining the MSA Transformer with structure information and combining the learned embeddings with a structure-based graph neural network (GNN) ( Jing et al, 2021b ), we show that we can obtain improved variant effect predictions when the input MSA is shallow. In contrast to other recent methods ( Chen et al, 2023 ), SSEmb is trained fully end-to-end, with a focus on integrating and aligning sequence- and structure-based information throughout the entire model. We show that resulting dual embedding of sequence and structure in SSEmb is information-rich, leads to accurate predictions of variant effects probed by MAVEs, and can be used for other downstream tasks besides variant effect prediction.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A joint embedding of protein sequence and structure enables robust variant effect predictions

Blaabjerg,

Jonsson,

Boomsma

et al. 2023

Preprint

View full text Add to dashboard Cite

The ability to predict how amino acid changes may affect protein function has a wide range of applications including in disease variant classification and protein engineering. Many existing methods focus on learning from patterns found in either protein sequences or protein structures. Here, we present a method for integrating information from protein sequences and structures in a single model that we term SSEmb (Sequence Structure Embedding). SSEmb combines a graph representation for the protein structure with a transformer model for processing multiple sequence alignments, and we show that by integrating both types of information we obtain a variant effect prediction model that is more robust to cases where sequence information is scarce. Furthermore, we find that SSEmb learns embeddings of the sequence and structural properties that are useful for other downstream tasks. We exemplify this by training a downstream model to predict protein-protein binding sites at high accuracy using only the SSEmb embeddings as input. We envisage that SSEmb may be useful both for zero-shot predictions of variant effects and as a representation for predicting protein properties that depend on protein sequence and structure.

show abstract

“…These vector representations capture the evolutionary and biophysical features of protein sequences in the representation model's latent embedding space 29,30 . Indeed, PLMs have been used to learn the structure of fitness landscapes [33][34][35] , and can learn evolutionary features when trained with ancestrally reconstructed sequence data 36 .…”

Section: Introductionmentioning

confidence: 99%

Ancestral reconstruction of polyethylene terephthalate degrading cutinases reveals a rugged and unexplored sequence-fitness landscape

Vongsouthi,

Georgelin,

Matthews

et al. 2024

Preprint

View full text Add to dashboard Cite

The enzymatic degradation of polyethylene terephthalate (PET) is a promising method of advanced plastic recycling. Traditional protein engineering methods often fall short in exploring protein sequence space for optimal enzymes due to structural and rational design limitations. Our study addresses this by using multiplexed ancestral sequence reconstruction (mASR) to explore the evolutionary sequence space of PET-degrading cutinases. With a dataset of 397 cutinase sequences, we created a diverse library of ancestral sequences. Experimental characterization of 48 ancestral sequences revealed a wide range of PETase activities, highlighting the value of mASR in uncovering functional variants when compared to traditional ASR. Our results showed that PETase activity in cutinases evolved through diverse pathways involving distal mutations to the active site, and is readily accessible within this family. Additionally, our analysis of the PETase fitness landscape using one-hot encoding (OHE) and local ancestral sequence embedding (LASE) highlighted the effectiveness of LASE in capturing sequence features relevant to activity. This work emphasizes the utility of mASR as a protein engineering tool for identifying enhanced PET-degrading enzymes, and the advantages of the LASE embedding scheme in mapping the PETase fitness landscape.

show abstract

Learning protein fitness landscapes with deep mutational scanning data from multiple sources

Cited by 4 publications

References 61 publications

Biophysics-based protein language models for protein engineering

Biophysics-based protein language models for protein engineering

A joint embedding of protein sequence and structure enables robust variant effect predictions

Ancestral reconstruction of polyethylene terephthalate degrading cutinases reveals a rugged and unexplored sequence-fitness landscape

Contact Info

Product

Resources

About