Search

The InterPro database (https://www.ebi.ac.uk/interpro/) provides an integrative classification of protein sequences into families, and identifies functionally important domains and conserved sites. Here, we report recent developments with InterPro (version 90.0) and its associated software, including updates to data content and to the website. These developments extend and enrich the information provided by InterPro, and provide a more user friendly access to the data. Additionally, we have worked on adding Pfam website features to the InterPro website, as the Pfam website will be retired in late 2022. We also show that InterPro's sequence coverage has kept pace with the growth of UniProtKB. Moreover, we report the development of a card game as a method of engaging the non-scientific community. Finally, we discuss the benefits and challenges brought by the use of artificial intelligence for protein structure prediction.

AlphaFold2 (AF2) has revolutionized structural biology by accurately predicting single structures of proteins and protein-protein complexes. However, biological function is rooted in a protein’s ability to sample different conformational substates, and disease-causing point mutations are often due to population changes of these substates. This has sparked immense interest in expanding AF2’s capability to predict conformational substates. We demonstrate that clustering an input multiple sequence alignment (MSA) by sequence similarity enables AF2 to sample alternate states of known metamorphic proteins, including the circadian rhythm protein KaiB, the transcription factor RfaH, and the spindle checkpoint protein Mad2, and score these states with high confidence. Moreover, we use AF2 to identify a minimal set of two point mutations predicted to switch KaiB between its two states. Finally, we used our clustering method, AF-cluster, to screen for alternate states in protein families without known fold-switching, and identified a putative alternate state for the oxidoreductase DsbE. Similarly to KaiB, DsbE is predicted to switch between a thioredoxin-like fold and a novel fold. This prediction is the subject of future experimental testing. Further development of such bioinformatic methods in tandem with experiments will likely have profound impact on predicting protein energy landscapes, essential for shedding light into biological function.

Multivariate chemical reaction optimization involving catalytic systems is a non-trivial task due to the high number of tuneable parameters and discrete choices. Closed-loop optimization featuring active Machine Learning (ML) represents a powerful strategy for automating reaction optimization. However, the translation of chemical reaction conditions into a machine-readable format comes with the challenge of finding highly informative features which accurately capture the factors for reaction success and allow the model to learn efficiently. Herein, we compare the efficacy of different calculated chemical descriptors for a high throughput generated dataset to determine the impact on a supervised ML model when predicting reaction yield. Then, the effect of featurization and size of the initial dataset within a closed-loop reaction optimization was examined. Finally, the balance between descriptor complexity and dataset size was considered. Ultimately, tailored descriptors did not outperform simple generic representations, however, a larger initial dataset accelerated reaction optimization.

The structured nature of chemical data means machine learning models trained to predict protein-ligand binding risk overfitting the data, impairing their ability to generalise and make accurate predictions for novel candidate ligands. To address this limitation, data debiasing algorithms systematically partition the data to reduce bias. When models are trained using debiased data splits, the reward for simply memorising the training data is reduced, suggesting that the ability of the model to make accurate predictions for novel candidate ligands will improve. To test this hypothesis, we use distance-based data splits to measure how well a model can generalise. We first confirm that models perform better for randomly split held-out sets than for distant held-out sets. We then debias the data and find, surprisingly, that debiasing typically reduces the ability of models to make accurate predictions for distant held-out test sets. These results suggest that debiasing reduces the information available to a model, impairing its ability to generalise.