Back

Benchmarking PROTAC docking and virtual screening tools

Rovers, E.; Schapira, M.

2023-09-01 bioinformatics
10.1101/2023.08.30.555318 bioRxiv
Show abstract

Proteolysis targeting chimeras (PROTACs) are bifunctional compounds that recruit an E3 ligase to a target protein to induce ubiquitination and degradation of the target and are pioneer molecules in the field of proximity pharmacology. Rational PROTAC design is a challenging process and novel computational tools have emerged that attempt to predict the ternary complexes created by PROTACs and identify PROTAC candidates. To compare the performance of recent PROTAC design and screening methods, a benchmark was developed to test the ability of these tools to 1) predict the ternary complexes observed in crystal structures and 2) dissociate active from inactive PROTACs. Unlike traditional protein-protein complex prediction software, the PROTAC virtual screening methods often generate successfully PROTAC-induced protein complex structures observed crystallographically, but these experimentally validated predictions are not dissociated from dozens or more of other predicted structures. PROTAC virtual screening efficiency is unclear and highly variable, in part due to the limited size of experimental datasets and the low number of negative controls. Defining ubiquitination zones within cullin-RING complexes does not improve predictions, but conformational arrangements can sometimes be found that are exclusively associated with active PROTACs. Computer assisted PROTAC design is still in its infancy. Pioneering tools highlight the promises and challenges in the field and may be more valuable when guided by clear structural and biophysical data and validated on specific chemical series.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.