BACKGROUND: Deep mutational scanning (DMS) provides comprehensive maps of protein variant effects but remains experimentally intensive. Machine learning (ML) approaches have the potential to reduce experimental burden of DMS by predicting the functional impact of substitutions from limited data. RESULTS: We introduced a ML classifier trained on normalised DMS scores from SARS-CoV-2 main protease (Mpro) to categorise amino acid substitutions as functional (wild-type-like) or non-functional. Using brute-force feature selection, we identified minimal subsets of six substitution scores per residue that enable accurate classification of the remaining substitutions, achieving minimum (worst accuracy) scores exceeding 90%. Models including support vector machines, random forests, and logistic regression were evaluated without retraining (zero-shot prediction) against additional SARS-CoV-2 Mpro datasets and against unrelated datasets. The zero-shot performance of the models was strongest for other enzymes and more modest when applied to DMS systems that assess protein folding and/or protein-protein interactions. CONCLUSION: The results show that targeted DMS combined with ML can reduce sequencing and reagent costs while preserving classification accuracy, offering a practical route to accelerate variant effect prediction.

Summary: MODE-TASK, a novel and versatile software suite, comprises Principal Component Analysis, Multidimensional Scaling, and t-Distributed Stochastic Neighbor Embedding techniques using Molecular Dynamics trajectories. MODE-TASK also includes a Normal Mode Analysis tool based on Anisotropic Network Model so as to provide a variety of ways to analyse and compare large-scale motions of protein complexes for which long MD simulations are prohibitive. Beside the command line function, a GUI has been developed as a PyMOL plugin. Availability and implementation: MODE-TASK is open source, and available for download from https://github.com/RUBi-ZA/MODE-TASK. It is implemented in Python and C++. It is compatible with Python 2.x and Python 3.x and can be installed by Conda. Supplementary information: Supplementary data are available at Bioinformatics online.

SUMMARY: Molecular dynamics (MD) determines the physical motions of atoms of a biological macromolecule in a cell-like environment and is an important method in structural bioinformatics. Traditionally, measurements such as root mean square deviation, root mean square fluctuation, radius of gyration, and various energy measures have been used to analyze MD simulations. Here, we present MD-TASK, a novel software suite that employs graph theory techniques, perturbation response scanning, and dynamic cross-correlation to provide unique ways for analyzing MD trajectories. AVAILABILITY AND IMPLEMENTATION: MD-TASK has been open-sourced and is available for download from https://github.com/RUBi-ZA/MD-TASK , implemented in Python and supported on Linux/Unix. CONTACT: [email protected].

Drug resistance in HIV treatment is still a worldwide problem. Predicting resistance to antiretrovirals (ARVs) before starting any treatment is important. Prediction accuracy is essential, as low-accuracy predictions increase the risk of prescribing sub-optimal drug regimens leading to patients developing resistance sooner. Artificial Neural Networks (ANNs) are a powerful tool that would be able to assist in drug resistance prediction. In this study, we constrained the dataset to subtype B, sacrificing generalizability for a higher predictive performance, and demonstrated that the predictive quality of the ANN regression models have definite improvement for most ARVs. Trained regression ANNs were optimized for eight protease inhibitors, six nucleoside reverse transcriptase (RT) inhibitors and four non-nucleoside RT inhibitors by experimenting combinations of rare variant filtering (none versus 1 residue occurrence) and ANN topologies (1–3 hidden layers with 2, 4, 6, 8 and 10 nodes per layer). Single hidden layers (5–20 nodes) were used for training where overfitting was detected. 5-fold cross-validation produced mean R 2 values over 0.95 and standard deviations lower than 0.04 for all but two antiretrovirals. Overall, higher accuracies and lower variances (compared to results published in 2016) were obtained by experimenting with various preprocessing methods, while focusing on the most prevalent subtype in the raw dataset (subtype B).We thus highlight the need to develop and make available subtype-specific datasets for developing higher accuracy in drug-resistance prediction methods.

Bioinformatics is now a critical skill in many research and commercial environments as biological data are increasing in both size and complexity. South African researchers recognized this need in the mid-1990s and responded by working with the government as well as international bodies to develop initiatives to build bioinformatics capacity in the country. Significant injections of support from these bodies provided a springboard for the establishment of computational biology units at multiple universities throughout the country, which took on teaching, basic research and support roles. Several challenges were encountered, for example with unreliability of funding, lack of skills, and lack of infrastructure. However, the bioinformatics community worked together to overcome these, and South Africa is now arguably the leading country in bioinformatics on the African continent. Here we discuss how the discipline developed in the country, highlighting the challenges, successes, and lessons learnt.

The discipline of bioinformatics has developed rapidly since the complete sequencing of the first genomes in the 1990s. The development of many high-throughput techniques during the last decades has ensured that bioinformatics has grown into a discipline that overlaps with, and is required for, the modern practice of virtually every field in the life sciences. This has placed a scientific premium on the availability of skilled bioinformaticians, a qualification that is extremely scarce on the African continent. The reasons for this are numerous, although the absence of a skilled bioinformatician at academic institutions to initiate a training process and build sustained capacity seems to be a common African shortcoming. This dearth of bioinformatics expertise has had a knock-on effect on the establishment of many modern high-throughput projects at African institutes, including the comprehensive and systematic analysis of genomes from African populations, which are among the most genetically diverse anywhere on the planet. Recent funding initiatives from the National Institutes of Health and the Wellcome Trust are aimed at ameliorating this shortcoming. In this paper, we discuss the problems that have limited the establishment of the bioinformatics field in Africa, as well as propose specific actions that will help with the education and training of bioinformaticians on the continent. This is an absolute requirement in anticipation of a boom in high-throughput approaches to human health issues unique to data from African populations.

The Research Unit in Bioinformatics at Rhodes University (RUBi), South Africa, offers a Masters of Science in Bioinformatics. Growing demand for bioinformatics qualifications results in applications from across Africa. Courses aim to bridge gaps in the diverse backgrounds of students who range from biologists with no prior computing exposure to computer scientists with no biology background. The programme is evenly split between coursework and research, with diverse modules from a range of departments covering mathematics, statistics, computer science and biology, with emphasis on application to bioinformatics research. The early focus on research helps bring students up to speed with working as a researcher. We measure success of the programme by the high rate of subsequent entry to PhD study: 10 of 14 students who completed in the years 2011-2013.