Predictive profiling of SARS-CoV-2 variants by deep mutational learning

Taft, J. M.; Weber, C. R.; Gao, B.; Ehling, R. A.; Han, J.; Frei, L.; Metcalfe, S. W.; Yermanos, A.; Kelton, W.; Reddy, S. T.

doi:10.1101/2021.12.07.471580

Home/Archive/bioRxiv

bioRxiv · 10.1101/2021.12.07.471580

Predictive profiling of SARS-CoV-2 variants by deep mutational learning

Taft, J. M.·Weber, C. R.·Gao, B.·Ehling, R. A.·Han, J.·Frei, L.·Metcalfe, S. W.·Yermanos, A.·Kelton, W.·Reddy, S. T.

View original source ↗Export metadata ↓

Abstract

The continual evolution of the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) and the emergence of variants that show resistance to vaccines and neutralizing antibodies (1-4) threaten to prolong the coronavirus disease 2019 (COVID-19) pandemic (5). Selection and emergence of SARS-CoV-2 variants are driven in part by mutations within the viral spike protein and in particular the ACE2 receptor-binding domain (RBD), a primary target site for neutralizing antibodies. Here, we develop deep mutational learning (DML), a machine learning-guided protein engineering technology, which is used to interrogate a massive sequence space of combinatorial mutations, representing billions of RBD variants, by accurately predicting their impact on ACE2 binding and antibody escape. A highly diverse landscape of possible SARS-CoV-2 variants is identified that could emerge from a multitude of evolutionary trajectories. DML may be used for predictive profiling on current and prospective variants, including highly mutated variants such as omicron (B.1.1.529), thus supporting decision making for public heath as well as guiding the development of therapeutic antibody treatments and vaccines for COVID-19.

Explore related subjects

synthetic biology Preprint

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Taft, J. M., Weber, C. R., Gao, B., Ehling, R. A., Han, J., Frei, L., Metcalfe, S. W., Yermanos, A., Kelton, W., Reddy, S. T.. 2021-12-09. Predictive profiling of SARS-CoV-2 variants by deep mutational learning. https://doi.org/10.1101/2021.12.07.471580

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Living electronic transistors with tunable conductivity

Electroactive bacteria, like Shewanella oneidensis, can couple the oxidation of organic electron donors to the reduction of external conductive surfaces, such as minerals and electrodes, by utilizing multiheme cytochromes to carry charge from within the cell to external surfaces. Additionally, multiheme cytochromes facilitate gateable, long-distance (micrometer-scale) redox conduction along the outer membrane and across multiple cells bridging electrodes. While electroactive microbes are being used to develop bioelectrochemical devices, there have been limited efforts to use synthetic biology to exert additional control over microbes serving as device components. Thus, this work implements an optogenetic biofilm patterning gene circuit and a small molecule sensor in S. oneidensis to simultaneously control cell deposition and cytochrome expression. This allows for photolithographic patterning of biofilms possessing tunable electrical properties controlled with small molecules. This system demonstrates tunable electrochemical activity, redox conduction, intrinsic biofilm conductivity, and negative differential transconductance as a function of cytochrome expression. Additionally, temperature-dependent measurements of this tunable biofilm conduction reveal changes in activation energy as a function of cytochrome expression. Through this combination of synthetic biology and electrochemistry, simultaneous control over biofilm geometry and conductivity sheds light on fundamental microbial electron transport processes, and it enables the construction of living electronic devices.

synthetic biology↗

bioRxiv2026

Evolutionary stabilisation of stressful metabolism via integrated biocomputing and essential-gene metabolic locking circuits

Synthetic genetic circuits enable microbial differentiation from growth to production, yet metabolic burden, imbalance and toxicity frequently drive strain degeneration. Yeast strains engineered to produce different terpene products exhibited divergent genetic responses to metabolic stresses, but commonly underwent progressive loss of induction of synthetic GAL regulatory circuits, either across the entire population or within subpopulations. Using di- and tri-input biocomputing circuits, the essential glutamine synthetase gene GLN1 was coupled to GAL induction, thereby enabling stabilisation and evolutionary adaptation of the synthetic genetic circuits and stressful heterologous terpene synthetic pathways. The integrated biocomputing and metabolic coupling circuit systems not only prevent strain degeneration but also enable interrogation of non-degenerative evolutionary shifts, providing a platform for metabolic engineering optimisation.

synthetic biology↗

bioRxiv2026

Unbiased and scalable reduction of diverse bacterial genomes

The genome is a complex, integrated system where the functions and regulatory interactions of its many components remain poorly understood. Genome minimization aims to reduce genomic complexity by removing non-essential elements to reveal the fundamental building blocks of cellular life. However, current minimization strategies are often slow and species-specific due to a reliance on prior information, and limited to producing single, isolated strains, which obscures the diverse ways a genome can adapt to large-scale DNA removal. Here we show the development and application of Stochastic Lineage-based Iterative Minimization (SLIM) a modular, high-throughput platform for unbiased genome reduction across phylogenetically diverse bacteria. We apply SLIM to generate a library of genome-reduced Escherichia coli lineages. We then interrogate the lineages, identifying both universal and lineage-specific transcriptional and translational reprogramming in response to deletions. We demonstrate that these expression dynamics drive environment-dependent fitness, allowing us to pinpoint a single gene deletion in one genome-reduced lineage as the driver of a measurable environmental growth defect. Beyond E. coli, we successfully deploy SLIM in phylogenetically distinct bacterial taxa to rapidly reduce the genomes of Shigella flexneri and Pseudomonas putida, distinct genus and order respectively from E. coli, without species-specific optimization. Our results establish a scalable, generalizable framework for navigating the vast landscape of minimized genomes, providing a powerful new tool for functional discovery and the rational design of synthetic genomic chassis.

synthetic biology↗