Large-scale protein language model improves prediction of harmful genetic mutations
In short: Researchers have developed a new computational workflow using a massive 650-million-parameter protein language model to predict the effects of nearly 450 million possible missense variants in the human genome. The study found that this approach outperformed existing methods at identifying whether known mutations are pathogenic or benign.
By applying the ESM1b model to clinical datasets and 28 deep mutational scan datasets, the researchers demonstrated that this method can also identify variants that only damage specific protein isoforms. While the approach generalizes to complex changes like stop-gains, its effectiveness relies on the patterns learned by the language model from protein sequences.
Why it matters: This scalable approach allows for the genome-wide assessment of how genetic changes impact protein function.
Opens the Intellect+ feed with this card first. Swipe up for the next one.
Sources
- Source: Research: Brandes et al. (2023). Genome-wide prediction of disease variant effects with a deep protein language model. Nature Genetics. · cc-by
To cite this card, link to https://intellectplus.co.uk/app/science-research-a-deep-learning-model-improves-prediction-of-how-genetic-mut.
How we source, check and correct cards: Editorial standards.
General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.
Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?
Related
More in Science · Latest · A–Z · All topics