Unsupervised Inference of Protein Fitness Landscape from Deep Mutational Scan

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.

Genotype sampling for deep-learning assisted experimental mapping of a combinatorially complete fitness landscape

MOTIVATION

Experimental characterization of fitness landscapes, which map genotypes onto fitness, is important for both evolutionary biology and protein engineering. It faces a fundamental obstacle in the astronomical number of genotypes whose fitness needs to be measured for any one protein. Deep learning may help to predict the fitness of many genotypes from a smaller neural network training sample of genotypes with experimentally measured fitness. Here I use a recently published experimentally mapped fitness landscape of more than 260 000 protein genotypes to ask how such sampling is best performed.

RESULTS

I show that multilayer perceptrons, recurrent neural networks, convolutional networks, and transformers, can explain more than 90% of fitness variance in the data. In addition, 90% of this performance is reached with a training sample comprising merely ≈103 sequences. Generalization to unseen test data is best when training data is sampled randomly and uniformly, or sampled to minimize the number of synonymous sequences. In contrast, sampling to maximize sequence diversity or codon usage bias reduces performance substantially. These observations hold for more than one network architecture. Simple sampling strategies may perform best when training deep learning neural networks to map fitness landscapes from experimental data.

AVAILABILITY AND IMPLEMENTATION

The fitness landscape data analyzed here is publicly available as described previously (Papkou et al. 2023). All code used to analyze this landscape is publicly available at https://github.com/andreas-wagner-uzh/fitness_landscape_sampling.

FLIGHTED: Inferring Fitness Landscapes from Noisy High-Throughput Experimental Data

Machine learning (ML) for protein design requires large protein fitness datasets generated by high-throughput experiments for training, fine-tuning, and benchmarking models. However, most models do not account for experimental noise inherent in these datasets, harming model performance and changing model rankings in benchmarking studies. Here, we develop FLIGHTED, a Bayesian method for generating fitness landscapes with calibrated errors from noisy high-throughput experimental data. We apply FLIGHTED to single-step selection assays such as phage display and to a novel high-throughput assay DHARMA that ties fitness to base editing activity. Our results show that FLIGHTED robustly generates fitness landscapes with accurate errors. We demonstrate that FLIGHTED improves model performance and enables the generation of protein fitness datasets of up to 106 variants with DHARMA. FLIGHTED can be used on any high-throughput assay and makes it easy for ML scientists to account for experimental noise when modeling protein fitness.

EvoAI enables extreme compression and reconstruction of the protein sequence space

Designing proteins with improved functions requires a deep understanding of how sequence and function are related, a vast space that is hard to explore. The ability to efficiently compress this space by identifying functionally important features is extremely valuable. Here, we first establish a method called EvoScan to comprehensively segment and scan the high-fitness sequence space to obtain anchor points that capture its essential features, especially in high dimensions. Our approach is compatible with any biomolecular function that can be coupled to a transcriptional output. We then develop deep learning and large language models to accurately reconstruct the space from these anchors, allowing computational prediction of novel, highly fit sequences without prior homology-derived or structural information. We apply this hybrid experimental-computational method, which we call EvoAI, to a repressor protein and find that only 82 anchors are sufficient to compress the high-fitness sequence space with a compression ratio of 1048. The extreme compressibility of the space informs both applied biomolecular design and understanding of natural evolution.

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.

Inferring protein fitness landscapes from laboratory evolution experiments

Directed laboratory evolution applies iterative rounds of mutation and selection to explore the protein fitness landscape and provides rich information regarding the underlying relationships between protein sequence, structure, and function. Laboratory evolution data consist of protein sequences sampled from evolving populations over multiple generations and this data type does not fit into established supervised and unsupervised machine learning approaches. We develop a statistical learning framework that models the evolutionary process and can infer the protein fitness landscape from multiple snapshots along an evolutionary trajectory. We apply our modeling approach to dihydrofolate reductase (DHFR) laboratory evolution data and the resulting landscape parameters capture important aspects of DHFR structure and function. We use the resulting model to understand the structure of the fitness landscape and find numerous examples of epistasis but an overall global peak that is evolutionarily accessible from most starting sequences. Finally, we use the model to perform an in silico extrapolation of the DHFR laboratory evolution trajectory and computationally design proteins from future evolutionary rounds.

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.

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.

Facilitating Machine Learning-Guided Protein Engineering with Smart Library Design and Massively Parallel Assays

Protein design plays an important role in recent medical advances from antibody therapy to vaccine design. Typically, exhaustive mutational screens or directed evolution experiments are used for the identification of the best design or for improvements to the wild-type variant. Even with a high-throughput screening on pooled libraries and Next-Generation Sequencing to boost the scale of read-outs, surveying all the variants with combinatorial mutations for their empirical fitness scores is still of magnitudes beyond the capacity of existing experimental settings. To tackle this challenge, in-silico approaches using machine learning to predict the fitness of novel variants based on a subset of empirical measurements are now employed. These machine learning models turn out to be useful in many cases, with the premise that the experimentally determined fitness scores and the amino-acid descriptors of the models are informative. The machine learning models can guide the search for the highest fitness variants, resolve complex epistatic relationships, and highlight bio-physical rules for protein folding. Using machine learning-guided approaches, researchers can build more focused libraries, thus relieving themselves from labor-intensive screens and fast-tracking the optimization process. Here, we describe the current advances in massive-scale variant screens, and how machine learning and mutagenesis strategies can be integrated to accelerate protein engineering. More specifically, we examine strategies to make screens more economical, informative, and effective in discovery of useful variants.

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.

Deep mutational scanning of the plasminogen activator inhibitor-1 functional landscape

The serine protease inhibitor (SERPIN) plasminogen activator inhibitor-1 (PAI-1) is a key regulator of the fibrinolytic system, inhibiting the serine proteases tissue- and urokinase-type plasminogen activator (tPA and uPA, respectively). Missense variants render PAI-1 non-functional through misfolding, leading to its turnover as a protease substrate, or to a more rapid transition to the latent/inactive state. Deep mutational scanning was performed to evaluate the impact of amino acid sequence variation on PAI-1 inhibition of uPA using an M13 filamentous phage display system. Error prone PCR was used to construct a mutagenized PAI-1 library encompassing ~ 70% of potential single amino acid substitutions. The relative effects of 27% of all possible missense variants on PAI-1 inhibition of uPA were determined using high-throughput DNA sequencing. 826 missense variants demonstrated conserved inhibitory activity while 1137 resulted in loss of PAI-1 inhibitory function. The least evolutionarily conserved regions of PAI-1 were also identified as being the most tolerant of missense mutations. The results of this screen confirm previous low-throughput mutational studies, including those of the reactive center loop. These data provide a powerful resource for explaining structure-function relationships for PAI-1 and for the interpretation of human genomic sequence variants.

Understanding the Origins of Loss of Protein Function by Analyzing the Effects of Thousands of Variants on Activity and Abundance

Understanding and predicting how amino acid substitutions affect proteins are keys to our basic understanding of protein function and evolution. Amino acid changes may affect protein function in a number of ways including direct perturbations of activity or indirect effects on protein folding and stability. We have analyzed 6,749 experimentally determined variant effects from multiplexed assays on abundance and activity in two proteins (NUDT15 and PTEN) to quantify these effects and find that a third of the variants cause loss of function, and about half of loss-of-function variants also have low cellular abundance. We analyze the structural and mechanistic origins of loss of function and use the experimental data to find residues important for enzymatic activity. We performed computational analyses of protein stability and evolutionary conservation and show how we may predict positions where variants cause loss of activity or abundance. In this way, our results link thermodynamic stability and evolutionary conservation to experimental studies of different properties of protein fitness landscapes.

Statistical Design of Experiments for Synthetic Biology

The design and optimization of biological systems is an inherently complex undertaking that requires careful balancing of myriad synergistic and antagonistic variables. However, despite this complexity, much synthetic biology research is predicated on One Factor at A Time (OFAT) experimentation; the genetic and environmental variables affecting the activity of a system of interest are sequentially altered while all other variables are held constant. Beyond being time and resource intensive, OFAT experimentation crucially ignores the effect of interactions between factors. Given the ubiquity of interacting genetic and environmental factors in biology this failure to account for interaction effects in OFAT experimentation can result in the development of suboptimal systems. To address these limitations, an increasing number of studies have turned to Design of Experiments (DoE), a suite of methods that enable efficient, systematic exploration and exploitation of complex design spaces. This review provides an overview of DoE for synthetic biologists. Key concepts and commonly used experimental designs are introduced, and we discuss the advantages of DoE as compared to OFAT experimentation. We dissect the applicability of DoE in the context of synthetic biology and review studies which have successfully employed these methods, illustrating the potential of statistical experimental design to guide the design, characterization, and optimization of biological protocols, pathways, and processes.