Abstract Motivation Protein solubility is a critical property that significantly impacts therapeutic efficacy and protein reengineering applications. Recent advances in machine learning and deep learning techniques provide unprecedented opportunities to develop predictive models for solubility, enabling more efficient protein design and optimization. This work is motivated by the potential of leveraging deep learning to address the solubility prediction challenge and to accelerate protein engineering workflows. Results Leveraging and combining multiple protein language model representations, our MiLaSol model attains 81% accuracy, outperforming prior methods, with the highest Matthews Correlation Coefficient (MCC) score of 0.63 demonstrating balanced performance across both soluble and insoluble proteins. Through simulated annealing optimization coupled with the Raygun model, we also present a computational method to reengineer insoluble protein variants into soluble forms, with predictions confirmed by multiple independent solubility prediction methods. Our results demonstrate the effectiveness of combining machine learning-based solubility prediction with generative optimization for protein engineering. Availability and implementation MiLaSol is available at https://github.com/weiweiloutufts/milasol An archived version of the code at the time of submission can be found at https://doi.org/10.5281/zenodo.21495202
Weiwei Lou, M. Erden, Lenore J. Cowen· Bioinformatics Advances· 0 citations
Nanobodies have recently emerged as an alternative to classical antibodies in therapeutic and diagnostic contexts, promising improved stability and simpler manufacturing. However, many labs still rely on low throughput conventional screening methods for nanobody discovery. Here we report streamlined experimental and computational tools for discovery of nanobodies, permitting deep characterization of the binding properties of immune repertoires.
To improve nanobody discovery, we developed NanoMAP, an integrated experimental and computational pipeline for nanobody discovery. We immunized alpacas with a pool of antigens, and created a phage display library from circulating B-cells. We then panned this phage display library on each antigen separately, and used competitors or antigen variants to assess complex binding phenotypes of the immune repertoire. Finally, we sequenced the panned libraries and developed a clustering method that allows data to be aggregated within B-cell clonal families, improving signal-to-noise ratios and reducing the complexity of the repertoire.
We tested NanoMAP on three distinct pools of targets, collecting data on close to 1M unique nanobody sequences. We found that our specialized clustering method outperformed standard sequence clustering, producing clonal families that are coherent in sequence and function.
By aggregating sequencing data within clonal families, NanoMAP produced reliable and rich data on binding phenotypes for each antigen. Using this information, we discovered nanobodies recognizing functionally relevant, and evolutionarily conserved sites on each antigen, demonstrating the broad utility of our methods.
NIAID R01AI25704, NIGMS 5K12GM133314-07
Computational and Systems Immunology (COMP)
William L. White, Edward H. Moseley, Jacqueline M. Tremblay et al.· Journal of Immunology· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
