A fundamental protein property, thermodynamic stability, revealed solely from large-scale measurements of protein function

Unsupervised modeling of mutational landscapes of adeno-associated viruses viability

Adeno-associated viruses 2 (AAV2) are minute viruses renowned for their capacity to infect human cells and akin organisms. They have recently emerged as prominent candidates in the field of gene therapy, primarily attributed to their inherent non-pathogenic nature in humans and the safety associated with their manipulation. The efficacy of AAV2 as gene therapy vectors hinges on their ability to infiltrate host cells, a phenomenon reliant on their competence to construct a capsid capable of breaching the nucleus of the target cell. To enhance their infection potential, researchers have extensively scrutinized various combinatorial libraries by introducing mutations into the capsid, aiming to boost their effectiveness. The emergence of high-throughput experimental techniques, like deep mutational scanning (DMS), has made it feasible to experimentally assess the fitness of these libraries for their intended purpose. Notably, machine learning is starting to demonstrate its potential in addressing predictions within the mutational landscape from sequence data. In this context, we introduce a biophysically-inspired model designed to predict the viability of genetic variants in DMS experiments. This model is tailored to a specific segment of the CAP region within AAV2's capsid protein. To evaluate its effectiveness, we conduct model training with diverse datasets, each tailored to explore different aspects of the mutational landscape influenced by the selection process. Our assessment of the biophysical model centers on two primary objectives: (i) providing quantitative forecasts for the log-selectivity of variants and (ii) deploying it as a binary classifier to categorize sequences into viable and non-viable classes.

Functional effects of mutations in proteins can be predicted and interpreted by guided selection of sequence covariation information

Predicting the effects of one or more mutations to the in vivo or in vitro properties of a wild-type protein is a major computational challenge, due to the presence of epistasis, that is, of interactions between amino acids in the sequence. We introduce a computationally efficient procedure to build minimal epistatic models to predict mutational effects by combining evolutionary (homologous sequence) and few mutational-scan data. Mutagenesis measurements guide the selection of links in a sparse graphical model, while the parameters on the nodes and the edges are inferred from sequence data. We show, on 10 mutational scans, that our pipeline exhibits performances comparable to state-of-the-art deep networks trained on many more data, while requiring much less parameters and being hence more interpretable. In particular, the identified interactions adapt to the wild-type protein and to the fitness or biochemical property experimentally measured, mostly focus on key functional sites, and are not necessarily related to structural contacts. Therefore, our method is able to extract information relevant for one mutational experiment from homologous sequence data reflecting the multitude of structural and functional constraints acting on proteins throughout evolution.

Epistasis facilitates functional evolution in an ancient transcription factor

A protein's genetic architecture - the set of causal rules by which its sequence produces its functions - also determines its possible evolutionary trajectories. Prior research has proposed that the genetic architecture of proteins is very complex, with pervasive epistatic interactions that constrain evolution and make function difficult to predict from sequence. Most of this work has analyzed only the direct paths between two proteins of interest - excluding the vast majority of possible genotypes and evolutionary trajectories - and has considered only a single protein function, leaving unaddressed the genetic architecture of functional specificity and its impact on the evolution of new functions. Here, we develop a new method based on ordinal logistic regression to directly characterize the global genetic determinants of multiple protein functions from 20-state combinatorial deep mutational scanning (DMS) experiments. We use it to dissect the genetic architecture and evolution of a transcription factor's specificity for DNA, using data from a combinatorial DMS of an ancient steroid hormone receptor's capacity to activate transcription from two biologically relevant DNA elements. We show that the genetic architecture of DNA recognition consists of a dense set of main and pairwise effects that involve virtually every possible amino acid state in the protein-DNA interface, but higher-order epistasis plays only a tiny role. Pairwise interactions enlarge the set of functional sequences and are the primary determinants of specificity for different DNA elements. They also massively expand the number of opportunities for single-residue mutations to switch specificity from one DNA target to another. By bringing variants with different functions close together in sequence space, pairwise epistasis therefore facilitates rather than constrains the evolution of new functions.

Network of epistatic interactions in an enzyme active site revealed by large-scale deep mutational scanning

Cooperative interactions between amino acids are critical for protein function. A genetic reflection of cooperativity is epistasis, which is when a change in the amino acid at one position changes the sequence requirements at another position. To assess epistasis within an enzyme active site, we utilized CTX-M β-lactamase as a model system. CTX-M hydrolyzes β-lactam antibiotics to provide antibiotic resistance, allowing a simple functional selection for rapid sorting of modified enzymes. We created all pairwise mutations across 17 active site positions in the β-lactamase enzyme and quantitated the function of variants against two β-lactam antibiotics using next-generation sequencing. Context-dependent sequence requirements were determined by comparing the antibiotic resistance function of double mutations across the CTX-M active site to their predicted function based on the constituent single mutations, revealing both positive epistasis (synergistic interactions) and negative epistasis (antagonistic interactions) between amino acid substitutions. The resulting trends demonstrate that positive epistasis is present throughout the active site, that epistasis between residues is mediated through substrate interactions, and that residues more tolerant to substitutions serve as generic compensators which are responsible for many cases of positive epistasis. Additionally, we show that a key catalytic residue (Glu166) is amenable to compensatory mutations, and we characterize one such double mutant (E166Y/N170G) that acts by an altered catalytic mechanism. These findings shed light on the unique biochemical factors that drive epistasis within an enzyme active site and will inform enzyme engineering efforts by bridging the gap between amino acid sequence and catalytic function.

Positive epistasis drives clavulanic acid resistance in double mutant libraries of BlaC β-lactamase

Phenotypic effects of mutations are highly dependent on the genetic backgrounds in which they occur, due to epistatic effects. To test how easily the loss of enzyme activity can be compensated for, we screen mutant libraries of BlaC, a β-lactamase from Mycobacterium tuberculosis, for fitness in the presence of carbenicillin and the inhibitor clavulanic acid. Using a semi-rational approach and deep sequencing, we prepare four double-site saturation libraries and determine the relative fitness effect for 1534/1540 (99.6%) of the unique library members at two temperatures. Each library comprises variants of a residue known to be relevant for clavulanic acid resistance as well as residue 105, which regulates access to the active site. Variants with greatly improved fitness were identified within each library, demonstrating that compensatory mutations for loss of activity can be readily found. In most cases, the fittest variants are a result of positive epistasis, indicating strong synergistic effects between the chosen residue pairs. Our study sheds light on a role of epistasis in the evolution of functional residues and underlines the highly adaptive potential of BlaC.

Identifying LasR Quorum Sensors with Improved Signal Specificity by Mapping the Sequence-Function Landscape

Programmable intercellular signaling using components of naturally occurring quorum sensing can allow for coordinated functions to be engineered in microbial consortia. LuxR-type transcriptional regulators are widely used for this purpose and are activated by homoserine lactone (HSL) signals. However, they often suffer from imperfect molecular discrimination of structurally similar HSLs, causing misregulation within engineered consortia containing multiple HSL signals. Here, we studied one such example, the regulator LasR from Pseudomonas aeruginosa. We elucidated its sequence-function relationship for ligand specificity using targeted protein engineering and multiplexed high-throughput biosensor screening. A pooled combinatorial saturation mutagenesis library (9,486 LasR DNA sequences) was created by mutating six residues in LasR's β5 sheet with single, double, or triple amino acid substitutions. Sort-seq assays were performed in parallel using cognate and noncognate HSLs to quantify each corresponding sensor's response to each HSL signal, which identified hundreds of highly specific variants. Sensor variants identified were individually assayed and exhibited up to 60.6-fold (p = 0.0013) improved relative activation by the cognate signal compared to the wildtype. Interestingly, we uncovered prevalent mutational epistasis and previously unidentified residues contributing to signal specificity. The resulting sensors with negligible signal crosstalk could be broadly applied to engineer bacteria consortia.

Inference of annealed protein fitness landscapes with AnnealDCA

The design of proteins with specific tasks is a major challenge in molecular biology with important diagnostic and therapeutic applications. High-throughput screening methods have been developed to systematically evaluate protein activity, but only a small fraction of possible protein variants can be tested using these techniques. Computational models that explore the sequence space in-silico to identify the fittest molecules for a given function are needed to overcome this limitation. In this article, we propose AnnealDCA, a machine-learning framework to learn the protein fitness landscape from sequencing data derived from a broad range of experiments that use selection and sequencing to quantify protein activity. We demonstrate the effectiveness of our method by applying it to antibody Rep-Seq data of immunized mice and screening experiments, assessing the quality of the fitness landscape reconstructions. Our method can be applied to several experimental cases where a population of protein variants undergoes various rounds of selection and sequencing, without relying on the computation of variants enrichment ratios, and thus can be used even in cases of disjoint sequence samples.

popDMS infers mutation effects from deep mutational scanning data

Deep mutational scanning (DMS) experiments provide a powerful method to measure the functional effects of genetic mutations at massive scales. However, the data generated from these experiments can be difficult to analyze, with significant variation between experimental replicates. To overcome this challenge, we developed popDMS, a computational method based on population genetics theory, to infer the functional effects of mutations from DMS data. Through extensive tests, we found that the functional effects of single mutations and epistasis inferred by popDMS are highly consistent across replicates, comparing favorably with existing methods. Our approach is flexible and can be widely applied to DMS data that includes multiple time points, multiple replicates, and different experimental conditions.

Selection pressures on evolution of ribonuclease H explored with rigorous free-energy-based design

Understanding natural protein evolution and designing novel proteins are motivating interest in development of high-throughput methods to explore large sequence spaces. In this work, we demonstrate the application of multisite λ dynamics (MSλD), a rigorous free energy simulation method, and chemical denaturation experiments to quantify evolutionary selection pressure from sequence-stability relationships and to address questions of design. This study examines a mesophilic phylogenetic clade of ribonuclease H (RNase H), furthering its extensive characterization in earlier studies, focusing on E. coli RNase H (ecRNH) and a more stable consensus sequence (AncCcons) differing at 15 positions. The stabilities of 32,768 chimeras between these two sequences were computed using the MSλD framework. The most stable and least stable chimeras were predicted and tested along with several other sequences, revealing a designed chimera with approximately the same stability increase as AncCcons, but requiring only half the mutations. Comparing the computed stabilities with experiment for 12 sequences reveals a Pearson correlation of 0.86 and root mean squared error of 1.18 kcal/mol, an unprecedented level of accuracy well beyond less rigorous computational design methods. We then quantified selection pressure using a simple evolutionary model in which sequences are selected according to the Boltzmann factor of their stability. Selection temperatures from 110 to 168 K are estimated in three ways by comparing experimental and computational results to evolutionary models. These estimates indicate selection pressure is high, which has implications for evolutionary dynamics and for the accuracy required for design, and suggests accurate high-throughput computational methods like MSλD may enable more effective protein design.

ACIDES: on-line monitoring of forward genetic screens for protein engineering

Forward genetic screens of mutated variants are a versatile strategy for protein engineering and investigation, which has been successfully applied to various studies like directed evolution (DE) and deep mutational scanning (DMS). While next-generation sequencing can track millions of variants during the screening rounds, the vast and noisy nature of the sequencing data impedes the estimation of the performance of individual variants. Here, we propose ACIDES that combines statistical inference and in-silico simulations to improve performance estimation in the library selection process by attributing accurate statistical scores to individual variants. We tested ACIDES first on a random-peptide-insertion experiment and then on multiple public datasets from DE and DMS studies. ACIDES allows experimentalists to reliably estimate variant performance on the fly and can aid protein engineering and research pipelines in a range of applications, including gene therapy.

ProteinGym: Large-Scale Benchmarks for Protein Design and Fitness Prediction

Predicting the effects of mutations in proteins is critical to many applications, from understanding genetic disease to designing novel proteins that can address our most pressing challenges in climate, agriculture and healthcare. Despite a surge in machine learning-based protein models to tackle these questions, an assessment of their respective benefits is challenging due to the use of distinct, often contrived, experimental datasets, and the variable performance of models across different protein families. Addressing these challenges requires scale. To that end we introduce ProteinGym, a large-scale and holistic set of benchmarks specifically designed for protein fitness prediction and design. It encompasses both a broad collection of over 250 standardized deep mutational scanning assays, spanning millions of mutated sequences, as well as curated clinical datasets providing high-quality expert annotations about mutation effects. We devise a robust evaluation framework that combines metrics for both fitness prediction and design, factors in known limitations of the underlying experimental methods, and covers both zero-shot and supervised settings. We report the performance of a diverse set of over 70 high-performing models from various subfields (eg., alignment-based, inverse folding) into a unified benchmark suite. We open source the corresponding codebase, datasets, MSAs, structures, model predictions and develop a user-friendly website that facilitates data access and analysis.

Ensemble Learning with Supervised Methods Based on Large-Scale Protein Language Models for Protein Mutation Effects Prediction

Machine learning has been increasingly utilized in the field of protein engineering, and research directed at predicting the effects of protein mutations has attracted increasing attention. Among them, so far, the best results have been achieved by related methods based on protein language models, which are trained on a large number of unlabeled protein sequences to capture the generally hidden evolutionary rules in protein sequences, and are therefore able to predict their fitness from protein sequences. Although numerous similar models and methods have been successfully employed in practical protein engineering processes, the majority of the studies have been limited to how to construct more complex language models to capture richer protein sequence feature information and utilize this feature information for unsupervised protein fitness prediction. There remains considerable untapped potential in these developed models, such as whether the prediction performance can be further improved by integrating different models to further improve the accuracy of prediction. Furthermore, how to utilize large-scale models for prediction methods of mutational effects on quantifiable properties of proteins due to the nonlinear relationship between protein fitness and the quantification of specific functionalities has yet to be explored thoroughly. In this study, we propose an ensemble learning approach for predicting mutational effects of proteins integrating protein sequence features extracted from multiple large protein language models, as well as evolutionarily coupled features extracted in homologous sequences, while comparing the differences between linear regression and deep learning models in mapping these features to quantifiable functional changes. We tested our approach on a dataset of 17 protein deep mutation scans and indicated that the integrated approach together with linear regression enables the models to have higher prediction accuracy and generalization. Moreover, we further illustrated the reliability of the integrated approach by exploring the differences in the predictive performance of the models across species and protein sequence lengths, as well as by visualizing clustering of ensemble and non-ensemble features.

Accurate prediction of functional effect of single amino acid variants with deep learning

The assessment of functional effect of amino acid variants is a critical biological problem in proteomics for clinical medicine and protein engineering. Although natively occurring variants offer insights into deleterious variants, high-throughput deep mutational experiments enable comprehensive investigation of amino acid variants for a given protein. However, these mutational experiments are too expensive to dissect millions of variants on thousands of proteins. Thus, computational approaches have been proposed, but they heavily rely on hand-crafted evolutionary conservation, limiting their accuracy. Recent advancement in transformers provides a promising solution to precisely estimate the functional effects of protein variants on high-throughput experimental data. Here, we introduce a novel deep learning model, namely Rep2Mut-V2, which leverages learned representation from transformer models. Rep2Mut-V2 significantly enhances the prediction accuracy for 27 types of measurements of functional effects of protein variants. In the evaluation of 38 protein datasets with 118,933 single amino acid variants, Rep2Mut-V2 achieved an average Spearman's correlation coefficient of 0.7. This surpasses the performance of six state-of-the-art methods, including the recently released methods ESM, DeepSequence and EVE. Even with limited training data, Rep2Mut-V2 outperforms ESM and DeepSequence, showing its potential to extend high-throughput experimental analysis for more protein variants to reduce experimental cost. In conclusion, Rep2Mut-V2 provides accurate predictions of the functional effects of single amino acid variants of protein coding sequences. This tool can significantly aid in the interpretation of variants in human disease studies.

Infer global, predict local: Quantity-relevance trade-off in protein fitness predictions from sequence data

Predicting the effects of mutations on protein function is an important issue in evolutionary biology and biomedical applications. Computational approaches, ranging from graphical models to deep-learning architectures, can capture the statistical properties of sequence data and predict the outcome of high-throughput mutagenesis experiments probing the fitness landscape around some wild-type protein. However, how the complexity of the models and the characteristics of the data combine to determine the predictive performance remains unclear. Here, based on a theoretical analysis of the prediction error, we propose descriptors of the sequence data, characterizing their quantity and relevance relative to the model. Our theoretical framework identifies a trade-off between these two quantities, and determines the optimal subset of data for the prediction task, showing that simple models can outperform complex ones when inferred from adequately-selected sequences. We also show how repeated subsampling of the sequence data is informative about how much epistasis in the fitness landscape is not captured by the computational model. Our approach is illustrated on several protein families, as well as on in silico solvable protein models.

Systematic profiling of dominant ubiquitin variants reveals key functional nodes contributing to evolutionary selection

Dominant-negative mutations can help to investigate the biological mechanisms and to understand the selective pressures for multifunctional proteins. However, most studies have focused on recessive mutant effects that occur in the absence of a second functional gene copy, which overlooks the fact that most eukaryotic genomes contain more than one copy of many genes. We have identified dominant effects on yeast growth rate among all possible point mutations in ubiquitin expressed alongside a wild-type allele. Our results reveal more than 400 dominant-negative mutations, indicating that dominant-negative effects make a sizable contribution to selection acting on ubiquitin. Cellular and biochemical analyses of individual ubiquitin variants show that dominant-negative effects are explained by varied accumulation of polyubiquitinated cellular proteins and/or defects in conjugation of ubiquitin variants to ubiquitin ligases. Our approach to identify dominant-negative mutations is general and can be applied to other proteins of interest.

A probabilistic view of protein stability, conformational specificity, and design

Various approaches have used neural networks as probabilistic models for the design of protein sequences. These "inverse folding" models employ different objective functions, which come with trade-offs that have not been assessed in detail before. This study introduces probabilistic definitions of protein stability and conformational specificity and demonstrates the relationship between these chemical properties and the [Formula: see text] Boltzmann probability objective. This links the Boltzmann probability objective function to experimentally verifiable outcomes. We propose a novel sequence decoding algorithm, referred to as "BayesDesign", that leverages Bayes' Rule to maximize the [Formula: see text] objective instead of the [Formula: see text] objective common in inverse folding models. The efficacy of BayesDesign is evaluated in the context of two protein model systems, the NanoLuc enzyme and the WW structural motif. Both BayesDesign and the baseline ProteinMPNN algorithm increase the thermostability of NanoLuc and increase the conformational specificity of WW. The possible sources of error in the model are analyzed.

Facile Method for High-throughput Identification of Stabilizing Mutations

Characterizing the effects of mutations on stability is critical for understanding the function and evolution of proteins and improving their biophysical properties. High throughput folding and abundance assays have been successfully used to characterize missense mutations associated with reduced stability. However, screening for increased thermodynamic stability is more challenging since such mutations are rarer and their impact on assay readout is more subtle. Here, a multiplex assay for high throughput screening of protein folding was developed by combining deep mutational scanning, fluorescence-activated cell sorting, and deep sequencing. By analyzing a library of 2000 variants of Adenylate kinase we demonstrate that the readout of the method correlates with stability and that mutants with up to 13 °C increase in thermal melting temperature could be identified with low false positive rate. The discovery of many stabilizing mutations also enabled the analysis of general substitution patterns associated with increased stability in Adenylate kinase. This high throughput method to identify stabilizing mutations can be combined with functional screens to identify mutations that improve both stability and activity.

Applications of artificial intelligence in clinical laboratory genomics

The transition from analog to digital technologies in clinical laboratory genomics is ushering in an era of "big data" in ways that will exceed human capacity to rapidly and reproducibly analyze those data using conventional approaches. Accurately evaluating complex molecular data to facilitate timely diagnosis and management of genomic disorders will require supportive artificial intelligence methods. These are already being introduced into clinical laboratory genomics to identify variants in DNA sequencing data, predict the effects of DNA variants on protein structure and function to inform clinical interpretation of pathogenicity, link phenotype ontologies to genetic variants identified through exome or genome sequencing to help clinicians reach diagnostic answers faster, correlate genomic data with tumor staging and treatment approaches, utilize natural language processing to identify critical published medical literature during analysis of genomic data, and use interactive chatbots to identify individuals who qualify for genetic testing or to provide pre-test and post-test education. With careful and ethical development and validation of artificial intelligence for clinical laboratory genomics, these advances are expected to significantly enhance the abilities of geneticists to translate complex data into clearly synthesized information for clinicians to use in managing the care of their patients at scale.

Protein engineering via sequence-performance mapping

Discovery and evolution of new and improved proteins has empowered molecular therapeutics, diagnostics, and industrial biotechnology. Discovery and evolution both require efficient screens and effective libraries, although they differ in their challenges because of the absence or presence, respectively, of an initial protein variant with the desired function. A host of high-throughput technologies-experimental and computational-enable efficient screens to identify performant protein variants. In partnership, an informed search of sequence space is needed to overcome the immensity, sparsity, and complexity of the sequence-performance landscape. Early in the historical trajectory of protein engineering, these elements aligned with distinct approaches to identify the most performant sequence: selection from large, randomized combinatorial libraries versus rational computational design. Substantial advances have now emerged from the synergy of these perspectives. Rational design of combinatorial libraries aids the experimental search of sequence space, and high-throughput, high-integrity experimental data inform computational design. At the core of the collaborative interface, efficient protein characterization (rather than mere selection of optimal variants) maps sequence-performance landscapes. Such quantitative maps elucidate the complex relationships between protein sequence and performance-e.g., binding, catalytic efficiency, biological activity, and developability-thereby advancing fundamental protein science and facilitating protein discovery and evolution.

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

One of the key points of machine learning-assisted directed evolution (MLDE) is the accurate learning of the fitness landscape, a conceptual mapping from sequence variants to the desired function. Here, we describe a multi-protein training scheme that leverages the existing deep mutational scanning data from diverse proteins to aid in understanding the fitness landscape of a new protein. Proof-of-concept trials are designed to validate this training scheme in three aspects: random and positional extrapolation for single-variant effects, zero-shot fitness predictions for new proteins, and extrapolation for higher-order variant effects from single-variant effects. Moreover, our study identified previously overlooked strong baselines, and their unexpectedly good performance brings our attention to the pitfalls of MLDE. Overall, these results may improve our understanding of the association between different protein fitness profiles and shed light on developing better machine learning-assisted approaches to the directed evolution of proteins. A record of this paper's transparent peer review process is included in the supplemental information.

Mega-scale experimental analysis of protein folding stability in biology and design

Advances in DNA sequencing and machine learning are providing insights into protein sequences and structures on an enormous scale1. However, the energetics driving folding are invisible in these structures and remain largely unknown2. The hidden thermodynamics of folding can drive disease3,4, shape protein evolution5-7 and guide protein engineering8-10, and new approaches are needed to reveal these thermodynamics for every sequence and structure. Here we present cDNA display proteolysis, a method for measuring thermodynamic folding stability for up to 900,000 protein domains in a one-week experiment. From 1.8 million measurements in total, we curated a set of around 776,000 high-quality folding stabilities covering all single amino acid variants and selected double mutants of 331 natural and 148 de novo designed protein domains 40-72 amino acids in length. Using this extensive dataset, we quantified (1) environmental factors influencing amino acid fitness, (2) thermodynamic couplings (including unexpected interactions) between protein sites, and (3) the global divergence between evolutionary amino acid usage and protein folding stability. We also examined how our approach could identify stability determinants in designed proteins and evaluate design methods. The cDNA display proteolysis method is fast, accurate and uniquely scalable, and promises to reveal the quantitative rules for how amino acid sequences encode folding stability.

An Atlas of Variant Effects to understand the genome at nucleotide resolution

Sequencing has revealed hundreds of millions of human genetic variants, and continued efforts will only add to this variant avalanche. Insufficient information exists to interpret the effects of most variants, limiting opportunities for precision medicine and comprehension of genome function. A solution lies in experimental assessment of the functional effect of variants, which can reveal their biological and clinical impact. However, variant effect assays have generally been undertaken reactively for individual variants only after and, in most cases long after, their first observation. Now, multiplexed assays of variant effect can characterise massive numbers of variants simultaneously, yielding variant effect maps that reveal the function of every possible single nucleotide change in a gene or regulatory element. Generating maps for every protein encoding gene and regulatory element in the human genome would create an 'Atlas' of variant effect maps and transform our understanding of genetics and usher in a new era of nucleotide-resolution functional knowledge of the genome. An Atlas would reveal the fundamental biology of the human genome, inform human evolution, empower the development and use of therapeutics and maximize the utility of genomics for diagnosing and treating disease. The Atlas of Variant Effects Alliance is an international collaborative group comprising hundreds of researchers, technologists and clinicians dedicated to realising an Atlas of Variant Effects to help deliver on the promise of genomics.

CovET: A covariation-evolutionary trace method that identifies protein structure-function modules

Measuring the relative effect that any two sequence positions have on each other may improve protein design or help better interpret coding variants. Current approaches use statistics and machine learning but rarely consider phylogenetic divergences which, as shown by Evolutionary Trace studies, provide insight into the functional impact of sequence perturbations. Here, we reframe covariation analyses in the Evolutionary Trace framework to measure the relative tolerance to perturbation of each residue pair during evolution. This approach (CovET) systematically accounts for phylogenetic divergences: at each divergence event, we penalize covariation patterns that belie evolutionary coupling. We find that while CovET approximates the performance of existing methods to predict individual structural contacts, it performs significantly better at finding structural clusters of coupled residues and ligand binding sites. For example, CovET found more functionally critical residues when we examined the RNA recognition motif and WW domains. It correlates better with large-scale epistasis screen data. In the dopamine D2 receptor, top CovET residue pairs recovered accurately the allosteric activation pathway characterized for Class A G protein-coupled receptors. These data suggest that CovET ranks highest the sequence position pairs that play critical functional roles through epistatic and allosteric interactions in evolutionarily relevant structure-function motifs. CovET complements current methods and may shed light on fundamental molecular mechanisms of protein structure and function.

The landscape of tolerated genetic variation in humans and primates

Personalized genome sequencing has revealed millions of genetic differences between individuals, but our understanding of their clinical relevance remains largely incomplete. To systematically decipher the effects of human genetic variants, we obtained whole-genome sequencing data for 809 individuals from 233 primate species and identified 4.3 million common protein-altering variants with orthologs in humans. We show that these variants can be inferred to have nondeleterious effects in humans based on their presence at high allele frequencies in other primate populations. We use this resource to classify 6% of all possible human protein-altering variants as likely benign and impute the pathogenicity of the remaining 94% of variants with deep learning, achieving state-of-the-art accuracy for diagnosing pathogenic variants in patients with genetic diseases.

Assessing variants of uncertain significance implicated in hearing loss using a comprehensive deafness proteome

Hearing loss is the leading sensory deficit, affecting ~ 5% of the population. It exhibits remarkable heterogeneity across 223 genes with 6328 pathogenic missense variants, making deafness-specific expertise a prerequisite for ascribing phenotypic consequences to genetic variants. Deafness-implicated variants are curated in the Deafness Variation Database (DVD) after classification by a genetic hearing loss expert panel and thorough informatics pipeline. However, seventy percent of the 128,167 missense variants in the DVD are "variants of uncertain significance" (VUS) due to insufficient evidence for classification. Here, we use the deep learning protein prediction algorithm, AlphaFold2, to curate structures for all DVD genes. We refine these structures with global optimization and the AMOEBA force field and use DDGun3D to predict folding free energy differences (∆∆GFold) for all DVD missense variants. We find that 5772 VUSs have a large, destabilizing ∆∆GFold that is consistent with pathogenic variants. When also filtered for CADD scores (> 25.7), we determine 3456 VUSs are likely pathogenic at a probability of 99.0%. Of the 224 genes in the DVD, 166 genes (74%) exhibit one or more missense variants predicted to cause a pathogenic change in protein folding stability. The VUSs prioritized here affect 119 patients (~ 3% of cases) sequenced by the OtoSCOPE targeted panel. Approximately half of these patients previously received an inconclusive report, and reclassification of these VUSs as pathogenic provides a new genetic diagnosis for six patients.

Marginal specificity in protein interactions constrains evolution of a paralogous family

The evolution of novel functions in biology relies heavily on gene duplication and divergence, creating large paralogous protein families. Selective pressure to avoid detrimental cross-talk often results in paralogs that exhibit exquisite specificity for their interaction partners. But how robust or sensitive is this specificity to mutation? Here, using deep mutational scanning, we demonstrate that a paralogous family of bacterial signaling proteins exhibits marginal specificity, such that many individual substitutions give rise to substantial cross-talk between normally insulated pathways. Our results indicate that sequence space is locally crowded despite overall sparseness, and we provide evidence that this crowding has constrained the evolution of bacterial signaling proteins. These findings underscore how evolution selects for "good enough" rather than optimized phenotypes, leading to restrictions on the subsequent evolution of paralogs.

The landscape of tolerated genetic variation in humans and primates

Personalized genome sequencing has revealed millions of genetic differences between individuals, but our understanding of their clinical relevance remains largely incomplete. To systematically decipher the effects of human genetic variants, we obtained whole genome sequencing data for 809 individuals from 233 primate species, and identified 4.3 million common protein-altering variants with orthologs in human. We show that these variants can be inferred to have non-deleterious effects in human based on their presence at high allele frequencies in other primate populations. We use this resource to classify 6% of all possible human protein-altering variants as likely benign and impute the pathogenicity of the remaining 94% of variants with deep learning, achieving state-of-the-art accuracy for diagnosing pathogenic variants in patients with genetic diseases.

One Sentence Summary

Deep learning classifier trained on 4.3 million common primate missense variants predicts variant pathogenicity in humans.

Widespread epistasis shapes RNA Polymerase II active site function and evolution

Multi-subunit RNA Polymerases (msRNAPs) are responsible for transcription in all kingdoms of life. At the heart of these msRNAPs is an ultra-conserved active site domain, the trigger loop (TL), coordinating transcription speed and fidelity by critical conformational changes impacting multiple steps in substrate selection, catalysis, and translocation. Previous studies have observed several different types of genetic interactions between eukaryotic RNA polymerase II (Pol II) TL residues, suggesting that the TL's function is shaped by functional interactions of residues within and around the TL. The extent of these interaction networks and how they control msRNAP function and evolution remain to be determined. Here we have dissected the Pol II TL interaction landscape by deep mutational scanning in Saccharomyces cerevisiae Pol II. Through analysis of over 15000 alleles, representing all single mutants, a rationally designed subset of double mutants, and evolutionarily observed TL haplotypes, we identify interaction networks controlling TL function. Substituting residues creates allele-specific networks and propagates epistatic effects across the Pol II active site. Furthermore, the interaction landscape further distinguishes alleles with similar growth phenotypes, suggesting increased resolution over the previously reported single mutant phenotypic landscape. Finally, co-evolutionary analyses reveal groups of co-evolving residues across Pol II converge onto the active site, where evolutionary constraints interface with pervasive epistasis. Our studies provide a powerful system to understand the plasticity of RNA polymerase mechanism and evolution, and provide the first example of pervasive epistatic landscape in a highly conserved and constrained domain within an essential enzyme.

Global Analysis of Multi-Mutants to Improve Protein Function

The identification of amino acid substitutions that both enhance the stability and function of a protein is a key challenge in protein engineering. Technological advances have enabled assaying thousands of protein variants in a single high-throughput experiment, and more recent studies use such data in protein engineering. We present a Global Multi-Mutant Analysis (GMMA) that exploits the presence of multiply-substituted variants to identify individual amino acid substitutions that are beneficial for the stability and function across a large library of protein variants. We have applied GMMA to a previously published experiment reporting on >54,000 variants of green fluorescent protein (GFP), each with known fluorescence output, and each carrying 1-15 amino acid substitutions (Sarkisyan et al., 2016). The GMMA method achieves a good fit to this dataset while being analytically transparent. We show experimentally that the six top-ranking substitutions progressively enhance GFP. More broadly, using only a single experiment as input our analysis recovers nearly all the substitutions previously reported to be beneficial for GFP folding and function. In conclusion, we suggest that large libraries of multiply-substituted variants may provide a unique source of information for protein engineering.

Assessing Variants of Uncertain Significance Implicated in Hearing Loss Using a Comprehensive Deafness Proteome

Hearing loss is the leading sensory deficit, affecting ~ 5% of the population. It exhibits remarkable heterogeneity across 223 genes with 6,328 pathogenic missense variants, making deafness-specific expertise a prerequisite for ascribing phenotypic consequences to genetic variants. Deafness-implicated variants are curated in the Deafness Variation Database (DVD) after classification by a genetic hearing loss expert panel and thorough informatics pipeline. However, seventy percent of the 128,167 missense variants in the DVD are "variants of uncertain significance" (VUS) due to insufficient evidence for classification. Here, we use the deep learning protein prediction algorithm, AlphaFold2, to curate structures for all DVD genes. We refine these structures with global optimization and the AMOEBA force field and use DDGun3D to predict folding free energy differences (∆∆G Fold ) for all DVD missense variants. We find that 5,772 VUSs have a large, destabilizing ∆∆G Fold that is consistent with pathogenic variants. When also filtered for CADD scores (> 25.7), we determine 3,456 VUSs are likely pathogenic at a probability of 99.0%. These VUSs affect 119 patients (~ 3% of cases) sequenced by the OtoSCOPE targeted panel. Approximately half of these patients previously received an inconclusive report, and reclassification of these VUSs as pathogenic provides a new genetic diagnosis for six patients.

Deep Antimicrobial Activity and Stability Analysis Inform Lysin Sequence-Function Mapping

Antibiotic-resistant infectious disease is a critical challenge to human health. Antimicrobial proteins offer a compelling solution if engineered for potency, selectivity, and physiological stability. Lysins, which lyse cells via degradation of cell wall peptidoglycans, have significant potential to fill this role. Yet, the functional complexity of antimicrobial activity has hindered high-throughput characterization for discovery and design. To dramatically expand knowledge of the sequence-function landscape of lysins, we developed a depletion-based assay for library-scale measurement of lysin inhibitory activity. We coupled this platform with a high-throughput proteolytic stability assay to assess the activity and stability of ∼5 × 104 lysin catalytic domain variants, resulting in the discovery of a variant with increased activity (70 ± 20%) and stability (7.2 ± 0.4 °C increased midpoint of thermal denaturation). Ridge regression of the resulting data set demonstrated that libraries with a higher average Hamming distance better informed pairwise models and that coupling activity and stability assays enabled better prediction of catalytically active lysins. The best models achieved Pearson's correlation coefficients of 0.87 ± 0.01 and 0.61 ± 0.04 for predicting catalytic domain stability and activity, respectively. Our work provides an efficient strategy for constructing protein sequence-function landscapes, drastically increases screening throughput for engineering lysins, and yields promising lysins for further development.

Deep mutational scanning: A versatile tool in systematically mapping genotypes to phenotypes

Unveiling how genetic variations lead to phenotypic variations is one of the key questions in evolutionary biology, genetics, and biomedical research. Deep mutational scanning (DMS) technology has allowed the mapping of tens of thousands of genetic variations to phenotypic variations efficiently and economically. Since its first systematic introduction about a decade ago, we have witnessed the use of deep mutational scanning in many research areas leading to scientific breakthroughs. Also, the methods in each step of deep mutational scanning have become much more versatile thanks to the oligo-synthesizing technology, high-throughput phenotyping methods and deep sequencing technology. However, each specific possible step of deep mutational scanning has its pros and cons, and some limitations still await further technological development. Here, we discuss recent scientific accomplishments achieved through the deep mutational scanning and describe widely used methods in each step of deep mutational scanning. We also compare these different methods and analyze their advantages and disadvantages, providing insight into how to design a deep mutational scanning study that best suits the aims of the readers' projects.

Proteome-Wide Structural Computations Provide Insights into Empirical Amino Acid Substitution Matrices

The relative contribution of mutation and selection to the amino acid substitution rates observed in empirical matrices is unclear. Herein, we present a neutral continuous fitness-stability model, inspired by the Arrhenius law (qij=aije-ΔΔGij). The model postulates that the rate of amino acid substitution (i→j) is determined by the product of a pre-exponential factor, which is influenced by the genetic code structure, and an exponential term reflecting the relative fitness of the amino acid substitutions. To assess the validity of our model, we computed changes in stability of 14,094 proteins, for which 137,073,638 in silico mutants were analyzed. These site-specific data were summarized into a 20 square matrix, whose entries, ΔΔGij, were obtained after averaging through all the sites in all the proteins. We found a significant positive correlation between these energy values and the disease-causing potential of each substitution, suggesting that the exponential term accurately summarizes the fitness effect. A remarkable observation was that amino acids that were highly destabilizing when acting as the source, tended to have little effect when acting as the destination, and vice versa (source → destination). The Arrhenius model accurately reproduced the pattern of substitution rates collected in the empirical matrices, suggesting a relevant role for the genetic code structure and a tuning role for purifying selection exerted via protein stability.

Increasing protein stability by inferring substitution effects from high-throughput experiments

We apply a computational model, global multi-mutant analysis (GMMA), to inform on effects of most amino acid substitutions from a randomly mutated gene library. Using a high mutation frequency, the method can determine mutations that increase the stability of even very stable proteins for which conventional selection systems have reached their limit. As a demonstration of this, we screened a mutant library of a highly stable and computationally redesigned model protein using an in vivo genetic sensor for folding and assigned a stability effect to 374 of 912 possible single amino acid substitutions. Combining the top 9 substitutions increased the unfolding energy 47 to 69 kJ/mol in a single engineering step. Crystal structures of stabilized variants showed small perturbations in helices 1 and 2, which rendered them closer in structure to the redesign template. This case study illustrates the capability of the method, which is applicable to any screen for protein function.

Intragenic compensation through the lens of deep mutational scanning

A significant fraction of mutations in proteins are deleterious and result in adverse consequences for protein function, stability, or interaction with other molecules. Intragenic compensation is a specific case of positive epistasis when a neutral missense mutation cancels effect of a deleterious mutation in the same protein. Permissive compensatory mutations facilitate protein evolution, since without them all sequences would be extremely conserved. Understanding compensatory mechanisms is an important scientific challenge at the intersection of protein biophysics and evolution. In human genetics, intragenic compensatory interactions are important since they may result in variable penetrance of pathogenic mutations or fixation of pathogenic human alleles in orthologous proteins from related species. The latter phenomenon complicates computational and clinical inference of an allele's pathogenicity. Deep mutational scanning is a relatively new technique that enables experimental studies of functional effects of thousands of mutations in proteins. We review the important aspects of the field and discuss existing limitations of current datasets. We reviewed ten published DMS datasets with quantified functional effects of single and double mutations and described rates and patterns of intragenic compensation in eight of them.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12551-022-01005-w.

A Simulation Analysis and Screening of Deleterious Nonsynonymous Single Nucleotide Polymorphisms (nsSNPs) in Sheep LEP Gene

Leptin is a polypeptide hormone produced in the adipose tissue and governs many processes in the body. Recently, polymorphisms in the LEP gene revealed a significant change in body weight regulation, energy balance, food intake, and reproductive hormone secretion. This study considers its crucial role in the regulation of the economically important traits of sheep. Several computational tools, including SIFT, Predict SNP2, SNAP2, and PROVEAN, have been used to screen out the deleterious nsSNPs. Following the screening of 11 nsSNPs in the sheep genome, 5 nsSNPs, T86M (C → T), D98N (G → A), N136T (A → C), R142Q (G → A), and P157Q (C → A), were predicted to have a significant deleterious effect on the LEP protein function, leading to phenotypic difference. The analysis of proteins’ stability change due to amino acid substitution using the I-stable, SDM, and DynaMut consistently confirmed that three nsSNPs (T86M (C → T), D98N (G → A), and P157Q (C → A)) increased protein stability. It is suggested that these three nsSNPs may enhance the evolvability of LEP protein, which is vital for the evolutionary adaptation of sheep. Our findings demonstrate that the five nsSNPs reported in this study might be responsible for sheep’s structural and functional modifications of LEP protein. This is the first comprehensive report on the sheep LEP gene. It narrow downs the candidate nsSNPs for in vitro experiments to facilitate the development of reliable molecular markers for associated traits.

Mechanistic insights into global suppressors of protein folding defects

Most amino acid substitutions in a protein either lead to partial loss-of-function or are near neutral. Several studies have shown the existence of second-site mutations that can rescue defects caused by diverse loss-of-function mutations. Such global suppressor mutations are key drivers of protein evolution. However, the mechanisms responsible for such suppression remain poorly understood. To address this, we characterized multiple suppressor mutations both in isolation and in combination with inactive mutants. We examined six global suppressors of the bacterial toxin CcdB, the known M182T global suppressor of TEM-1 β-lactamase, the N239Y global suppressor of p53-DBD and three suppressors of the SARS-CoV-2 spike Receptor Binding Domain. When coupled to inactive mutants, they promote increased in-vivo solubilities as well as regain-of-function phenotypes. In the case of CcdB, where novel suppressors were isolated, we determined the crystal structures of three such suppressors to obtain insight into the specific molecular interactions responsible for the observed effects. While most individual suppressors result in small stability enhancements relative to wildtype, which can be combined to yield significant stability increments, thermodynamic stabilisation is neither necessary nor sufficient for suppressor action. Instead, in diverse systems, we observe that individual global suppressors greatly enhance the foldability of buried site mutants, primarily through increase in refolding rate parameters measured in vitro. In the crowded intracellular environment, mutations that slow down folding likely facilitate off-pathway aggregation. We suggest that suppressor mutations that accelerate refolding can counteract this, enhancing the yield of properly folded, functional protein in vivo.

Sequence dependencies and biophysical features both govern cleavage of diverse cut-sites by HIV protease

The infectivity of HIV-1 requires its protease (PR) cleave multiple cut-sites with low sequence similarity. The diversity of cleavage sites has made it challenging to investigate the underlying sequence properties that determine binding and turnover of substrates by PR. We engineered a mutational scanning approach utilizing yeast display, flow cytometry, and deep sequencing to systematically measure the impacts of all individual amino acid changes at 12 positions in three different cut-sites (MA/CA, NC/p1, and p1/p6). The resulting fitness landscapes revealed common physical features that underlie cutting of all three cut-sites at the amino acid positions closest to the scissile bond. In contrast, positions more than two amino acids away from the scissile bond exhibited a strong dependence on the sequence background of the rest of the cut-site. We observed multiple amino acid changes in cut-sites that led to faster cleavage rates, including a preference for negative charge five and six amino acids away from the scissile bond at locations where the surface of protease is positively charged. Analysis of individual cut sites using full-length matrix-capsid proteins indicate that long-distance sequence context can contribute to cutting efficiency such that analyses of peptides or shorter engineered constructs including those in this work should be considered carefully. This work provides a framework for understanding how diverse substrates interact with HIV-1 PR and can be extended to investigate other viral PRs with similar properties.

Interpretable pairwise distillations for generative protein sequence models

Many different types of generative models for protein sequences have been proposed in literature. Their uses include the prediction of mutational effects, protein design and the prediction of structural properties. Neural network (NN) architectures have shown great performances, commonly attributed to the capacity to extract non-trivial higher-order interactions from the data. In this work, we analyze two different NN models and assess how close they are to simple pairwise distributions, which have been used in the past for similar problems. We present an approach for extracting pairwise models from more complex ones using an energy-based modeling framework. We show that for the tested models the extracted pairwise models can replicate the energies of the original models and are also close in performance in tasks like mutational effect prediction. In addition, we show that even simpler, factorized models often come close in performance to the original models.

Complex neural networks trained on large biological datasets have recently shown powerful capabilites in tasks like the prediction of protein structure, assessing the effect of mutations on the fitness of proteins and even designing completely novel proteins with desired characteristics. The enthralling prospect of leveraging these advances in fields like medicine and synthetic biology has created a large amount of interest in academic research and industry. The connected question of what biological insights these methods actually gain during training has, however, received less attention. In this work, we systematically investigate in how far neural networks capture information that could not be captured by simpler models. To this end, we develop a method to train simpler models to imitate more complex models, and compare their performance to the original neural network models. Surprisingly, we find that the simpler models thus trained often perform on par with the neural networks, while having a considerably easier structure. This highlights the importance of finding ways to interpret the predictions of neural networks in these fields, which could inform the creation of better models, improve methods for their assessment and ultimately also increase our understanding of the underlying biology.

From systems to structure - using genetic data to model protein structures

Understanding the effects of genetic variation is a fundamental problem in biology that requires methods to analyse both physical and functional consequences of sequence changes at systems-wide and mechanistic scales. To achieve a systems view, protein interaction networks map which proteins physically interact, while genetic interaction networks inform on the phenotypic consequences of perturbing these protein interactions. Until recently, understanding the molecular mechanisms that underlie these interactions often required biophysical methods to determine the structures of the proteins involved. The past decade has seen the emergence of new approaches based on coevolution, deep mutational scanning and genome-scale genetic or chemical-genetic interaction mapping that enable modelling of the structures of individual proteins or protein complexes. Here, we review the emerging use of large-scale genetic datasets and deep learning approaches to model protein structures and their interactions, and discuss the integration of structural data from different sources.

Environmental selection and epistasis in an empirical phenotype-environment-fitness landscape

Fitness landscapes, mappings of genotype/phenotype to their effects on fitness, are invaluable concepts in evolutionary biochemistry. Although widely discussed, measurements of phenotype-fitness landscapes in proteins remain scarce. Here, we quantify all single mutational effects on fitness and phenotype (EC₅₀) of VIM-2 β-lactamase across a 64-fold range of ampicillin concentrations. We then construct a phenotype-fitness landscape that takes variations in environmental selection pressure into account. We found that a simple, empirical landscape accurately models the ~39,000 mutational data points, suggesting that the evolution of VIM-2 can be predicted on the basis of the selection environment. Our landscape provides new quantitative knowledge on the evolution of the β-lactamases and proteins in general, particularly their evolutionary dynamics under subinhibitory antibiotic concentrations, as well as the mechanisms and environmental dependence of non-specific epistasis.

Identification of stabilizing point mutations through mutagenesis of destabilized protein libraries

Although there have been recent transformative advances in the area of protein structure prediction, prediction of point mutations that improve protein stability remains challenging. It is possible to construct and screen large mutant libraries for improved activity or ligand binding. However, reliable screens for mutants that improve protein stability do not yet exist, especially for proteins that are well folded and relatively stable. Here, we demonstrate that incorporation of a single, specific, destabilizing mutation termed parent inactivating mutation into each member of a single-site saturation mutagenesis library, followed by screening for suppressors, allows for robust and accurate identification of stabilizing mutations. We carried out fluorescence-activated cell sorting of such a yeast surface display, saturation suppressor library of the bacterial toxin CcdB, followed by deep sequencing of sorted populations. We found that multiple stabilizing mutations could be identified after a single round of sorting. In addition, multiple libraries with different parent inactivating mutations could be pooled and simultaneously screened to further enhance the accuracy of identification of stabilizing mutations. Finally, we show that individual stabilizing mutations could be combined to result in a multi-mutant that demonstrated an increase in thermal melting temperature of about 20 °C, and that displayed enhanced tolerance to high temperature exposure. We conclude that as this method is robust and employs small library sizes, it can be readily extended to other display and screening formats to rapidly isolate stabilized protein mutants.

Learning protein fitness models from evolutionary and assay-labeled data

Machine learning-based models of protein fitness typically learn from either unlabeled, evolutionarily related sequences or variant sequences with experimentally measured labels. For regimes where only limited experimental data are available, recent work has suggested methods for combining both sources of information. Toward that goal, we propose a simple combination approach that is competitive with, and on average outperforms more sophisticated methods. Our approach uses ridge regression on site-specific amino acid features combined with one probability density feature from modeling the evolutionary data. Within this approach, we find that a variational autoencoder-based probability density model showed the best overall performance, although any evolutionary density model can be used. Moreover, our analysis highlights the importance of systematic evaluations and sufficient baselines.

Determinants of trafficking, conduction, and disease within a K+ channel revealed through multiparametric deep mutational scanning

A long-standing goal in protein science and clinical genetics is to develop quantitative models of sequence, structure, and function relationships to understand how mutations cause disease. Deep mutational scanning (DMS) is a promising strategy to map how amino acids contribute to protein structure and function and to advance clinical variant interpretation. Here, we introduce 7429 single-residue missense mutations into the inward rectifier K⁺ channel Kir2.1 and determine how this affects folding, assembly, and trafficking, as well as regulation by allosteric ligands and ion conduction. Our data provide high-resolution information on a cotranslationally folded biogenic unit, trafficking and quality control signals, and segregated roles of different structural elements in fold stability and function. We show that Kir2.1 surface trafficking mutants are underrepresented in variant effect databases, which has implications for clinical practice. By comparing fitness scores with expert-reviewed variant effects, we can predict the pathogenicity of 'variants of unknown significance' and disease mechanisms of known pathogenic mutations. Our study in Kir2.1 provides a blueprint for how multiparametric DMS can help us understand the mechanistic basis of genetic disorders and the structure-function relationships of proteins.

Deep mutational scanning for therapeutic antibody engineering

The biophysical and functional properties of monoclonal antibody (mAb) drug candidates are often improved by protein engineering methods to increase the probability of clinical efficacy. One emerging method is deep mutational scanning (DMS) which combines the power of exhaustive protein mutagenesis and functional screening with deep sequencing and bioinformatics. The application of DMS has yielded significant improvements to the affinity, specificity, and stability of several preclinical antibodies alongside novel applications such as introducing multi-specific binding properties. DMS has also been applied directly on target antigens to precisely map antibody-binding epitopes and notably to profile the mutational escape potential of viral targets (e.g., SARS-CoV-2 variants). Finally, DMS combined with machine learning is enabling advances in the computational screening and engineering of therapeutic antibodies.

A dual-reporter system for investigating and optimizing protein translation and folding in E. coli

Strategies for investigating and optimizing the expression and folding of proteins for biotechnological and pharmaceutical purposes are in high demand. Here, we describe a dual-reporter biosensor system that simultaneously assesses in vivo protein translation and protein folding, thereby enabling rapid screening of mutant libraries. We have validated the dual-reporter system on five different proteins and find an excellent correlation between reporter signals and the levels of protein expression and solubility of the proteins. We further demonstrate the applicability of the dual-reporter system as a screening assay for deep mutational scanning experiments. The system enables high throughput selection of protein variants with high expression levels and altered protein stability. Next generation sequencing analysis of the resulting libraries of protein variants show a good correlation between computationally predicted and experimentally determined protein stabilities. We furthermore show that the mutational experimental data obtained using this system may be useful for protein structure calculations.

AMaLa: Analysis of Directed Evolution Experiments via Annealed Mutational Approximated Landscape

We present Annealed Mutational approximated Landscape (AMaLa), a new method to infer fitness landscapes from Directed Evolution experiments sequencing data. Such experiments typically start from a single wild-type sequence, which undergoes Darwinian in vitro evolution via multiple rounds of mutation and selection for a target phenotype. In the last years, Directed Evolution is emerging as a powerful instrument to probe fitness landscapes under controlled experimental conditions and as a relevant testing ground to develop accurate statistical models and inference algorithms (thanks to high-throughput screening and sequencing). Fitness landscape modeling either uses the enrichment of variants abundances as input, thus requiring the observation of the same variants at different rounds or assuming the last sequenced round as being sampled from an equilibrium distribution. AMaLa aims at effectively leveraging the information encoded in the whole time evolution. To do so, while assuming statistical sampling independence between sequenced rounds, the possible trajectories in sequence space are gauged with a time-dependent statistical weight consisting of two contributions: (i) an energy term accounting for the selection process and (ii) a generalized Jukes-Cantor model for the purely mutational step. This simple scheme enables accurately describing the Directed Evolution dynamics and inferring a fitness landscape that correctly reproduces the measures of the phenotype under selection (e.g., antibiotic drug resistance), notably outperforming widely used inference strategies. In addition, we assess the reliability of AMaLa by showing how the inferred statistical model could be used to predict relevant structural properties of the wild-type sequence.

ECNet is an evolutionary context-integrated deep learning framework for protein engineering

Machine learning has been increasingly used for protein engineering. However, because the general sequence contexts they capture are not specific to the protein being engineered, the accuracy of existing machine learning algorithms is rather limited. Here, we report ECNet (evolutionary context-integrated neural network), a deep-learning algorithm that exploits evolutionary contexts to predict functional fitness for protein engineering. This algorithm integrates local evolutionary context from homologous sequences that explicitly model residue-residue epistasis for the protein of interest with the global evolutionary context that encodes rich semantic and structural features from the enormous protein sequence universe. As such, it enables accurate mapping from sequence to function and provides generalization from low-order mutants to higher-order mutants. We show that ECNet predicts the sequence-function relationship more accurately as compared to existing machine learning algorithms by using ~50 deep mutational scanning and random mutagenesis datasets. Moreover, we used ECNet to guide the engineering of TEM-1 β-lactamase and identified variants with improved ampicillin resistance with high success rates.

Saturation Mutagenesis of the Transmembrane Region of HokC in Escherichia coli Reveals Its High Tolerance to Mutations

Cells adapt to different stress conditions, such as the antibiotics presence. This adaptation sometimes is achieved by changing relevant protein positions, of which the mutability is limited by structural constrains. Understanding the basis of these constrains represent an important challenge for both basic science and potential biotechnological applications. To study these constraints, we performed a systematic saturation mutagenesis of the transmembrane region of HokC, a toxin used by Escherichia coli to control its own population, and observed that 92% of single-point mutations are tolerated and that all the non-tolerated mutations have compensatory mutations that reverse their effect. We provide experimental evidence that HokC accumulates multiple compensatory mutations that are found as correlated mutations in the HokC family multiple sequence alignment. In agreement with these observations, transmembrane proteins show higher probability to present correlated mutations and are less densely packed locally than globular proteins; previous mutagenesis results on transmembrane proteins further support our observations on the high tolerability to mutations of transmembrane regions of proteins. Thus, our experimental results reveal the HokC transmembrane region high tolerance to loss-of-function mutations that is associated with low sequence conservation and high rate of correlated mutations in the HokC family sequences alignment, which are features shared with other transmembrane proteins.