University of Glasgow

Brain MRI morphometry framework turning one T1w scan into 13 descriptors and a normative deviation profile shared across 19 clinical diagnoses.

Structural brain MRI foundation model pretrained on synthetic healthy volumes only, giving frozen features for brain age, dementia risk and image QC.

Protein language models evotuned on influenza A hemagglutinin, with a pLM entropy metric scoring per-site conservation from a single input sequence.

Protein-protein interaction prediction from a language model that encodes both sequences jointly, trained on human PPIs and applied across species.

Brain MRI segmentation network with progressive levels of detail, trained across ~160 acquisition sites so one checkpoint handles unseen scanners.

Brain MRI segmentation model that labels an entire 7T T1w volume in a single pass, returning six tissue classes plus background in seconds.