Computational pipeline
BE-PROTECT creates enhanced 'made in Belgium' biodefence measures aimed at optimising assessment, prevention, and preparedness, as well as response and recovery in the event of biological incidents. To do so, the project formulates a computational pipeline for extensive in silico protein characterisation that interprets experimental genome-level information on bioagents, naturally occurring or released with malicious intent.
The crucial novelty in the approach is that it also accounts for the non-folding flexible regions of proteins. Many pathogen mechanisms rely on such flexible regions to infect and interfere with host cells, but because of the difficulty in characterising and interpreting these regions, they are mostly ignored.
BE-PROTECT achieves this goal by combining publicly available data, available methods and responsible use of AI to identify the function and likely targets of bioagent proteins, compare them with annotated pathogen proteins and with host proteins, and then identify likely protein interactions and possible bioagent disruptor drugs through drug repurposing and AI-designed mini-binders.
The computational pipeline is intended to be fully dual-use with a public component for general public health measures and a private component for biodefence relating to sensitive bioagents, where the input data for the pipeline and its results can be handled securely in a fully parallel fashion.