. 2025 Feb 24;65(5):2304–2313. doi: 10.1021/acs.jcim.4c01752

Table 3. Performance of Predicting Thermostability with Regression Tasks Using the Dataset in DeepStabP^a.

method	R²	Spearman	Pearson	model parameter
DeepStabP (ProtT5-XL)⁴⁸ (reproduced)	0.792 ± 0.021	0.732 ± 0.005	0.897 ± 0.003	3B
ProCeSa-ESM-C (+Cont)	0.788 ± 0.004	0.717 ± 0.004	0.889 ± 0.004	600M

R², Spearman, and Pearson correlation coefficients are reported. The approximate number of model parameters is listed for both models.

Table 3. Performance of Predicting Thermostability with Regression Tasks Using the Dataset in DeepStabPa.