Fig. 3: FusOn-pLM embedding benchmarks on puncta prediction tasks.
From: FusOn-pLM: a fusion oncoprotein-specific language model via adjusted rate masking
A Certain FOs form puncta (condensates) via phase separation. Puncta may localize to the nucleus, cytoplasm, or both. B Three XGBoost classifiers are trained on FusOn-pLM-embedded FOs. One predicts formation of puncta (puncta propensity); one predicts formation of nuclear puncta (nucleus localization); one predicts formation of cytoplasmic puncta (cytoplasm localization). C–E Performance on a held-out test set when predictors are trained on FusOn-pLM, ESM-2-650M, ProtT5-XL-U50, and FOdb embeddings. Created in BioRender. Chatterjee, P. (2025) https://BioRender.com/e57u556. Source data for this figure are provided in the Source Data file.
