Journal: bioRxiv

Article Title: Structure-aware geometric graph learning for modeling protease–substrate specificity at scale

doi: 10.64898/2026.04.08.717168

Figure Lengend Snippet: a. Experimental workflow: substrate proteins were prepared via E . coli expression or commercial purification, followed by in vitro cleavage assays (control vs. experimental), SDS-PAGE, and LC-MS/MS analysis. b . SDS-PAGE results show markedly reduced bands for THOC5, RPIA, and CUL7 in experimental groups, indicating cleavage. c . Cleavage sites detected in vitro are compared with predictions from OmniCleave and Procleave; Seqlogo plots illustrate P4–P1 sequence preferences of validated substrates and three candidate proteins. d . Cleavage site distribution across protein domains: THOC5 sites localise mainly to the Tandem RWD domain, while CUL7 and RPIA sites are predominantly in non-domain regions. e . Molecular mechanism of Caspase-3 cleavage: hydrogen bonds, salt bridges, and hydrophobic interactions between Caspase-3 catalytic residues (H121 and C163) and key substrate residues (e.g., CUL7-D318, THOC5-D306) reveal the structural basis for substrate recognition and catalysis.

Article Snippet: Full-length recombinant human Caspase-3 (CASP3, UniprotKB: P42574) and RPIA (UniprotKB: P49247) used in this study were purchased commercially (CASP3: MedChemExpress #HY-P701341; RPIA: FinTest #P5260).

Techniques: Expressing, Purification, In Vitro, Control, SDS Page, Liquid Chromatography with Mass Spectroscopy, Sequencing

Journal: bioRxiv

Article Title: Structure-aware geometric graph learning for modeling protease–substrate specificity at scale

doi: 10.64898/2026.04.08.717168

Figure Lengend Snippet: a. Experimental workflow: substrate proteins were prepared via E . coli expression or commercial purification, followed by in vitro cleavage assays (control vs. experimental), SDS-PAGE, and LC-MS/MS analysis. b . SDS-PAGE results show markedly reduced bands for THOC5, RPIA, and CUL7 in experimental groups, indicating cleavage. c . Cleavage sites detected in vitro are compared with predictions from OmniCleave and Procleave; Seqlogo plots illustrate P4–P1 sequence preferences of validated substrates and three candidate proteins. d . Cleavage site distribution across protein domains: THOC5 sites localise mainly to the Tandem RWD domain, while CUL7 and RPIA sites are predominantly in non-domain regions. e . Molecular mechanism of Caspase-3 cleavage: hydrogen bonds, salt bridges, and hydrophobic interactions between Caspase-3 catalytic residues (H121 and C163) and key substrate residues (e.g., CUL7-D318, THOC5-D306) reveal the structural basis for substrate recognition and catalysis.

Article Snippet: Full-length recombinant human Caspase-3 (CASP3, UniprotKB: P42574) and RPIA (UniprotKB: P49247) used in this study were purchased commercially (CASP3: MedChemExpress #HY-P701341; RPIA: FinTest #P5260).

Techniques: Expressing, Purification, In Vitro, Control, SDS Page, Liquid Chromatography with Mass Spectroscopy, Sequencing

Journal: Annals of the New York Academy of Sciences

Article Title: Reassessing Transketolase Assays: Methodological Considerations for Detecting Functional Thiamine Deficiency

doi: 10.1111/nyas.70245

Figure Lengend Snippet: An overview of the pentose‐phosphate pathway (PPP). In the oxidative phase of the PPP, glucose‐6‐phosphate (G6P) is irreversibly converted to 6‐phosphogluconolactone (6PGL) by glucose‐6‐phosphate dehydrogenase (G6PD), then to 6‐phosphogluconate (6PG) by 6‐phosphonogluconolactonase, and finally to ribulose‐5‐phosphate (Ru5P) by 6‐phosphogluconate dehydrogenase (6PGD), generating NADPH. In the nonoxidative phase, Ru5P is converted to ribose‐5‐phosphate (R5P) by ribose‐5‐phosphate isomerase (RPIA) or to xylulose‐5‐phosphate (X5P) by ribose‐5‐phosphate epimerase (RPE). Transketolase (TKT) catalyzes thiamine diphosphate (TDP)‐dependent carbon transfers in the nonoxidative phase, converting R5P and X5P to sedoheptulose‐7‐phosphate (S7P) and glyceraldehyde‐3‐phosphate (G3P), and erythrose‐4‐phosphate (E4P) and X5P to fructose‐6‐phosphate (F6P) and G3P, respectively. These products can feed back into glycolysis directly, or F6P can be converted back to G6P using glucose‐6‐phosphate isomerase (GPI) to feed back into either glycolysis or the oxidative phase of the PPP.

Article Snippet: Per the Human Protein Atlas, both RPIA and RPE exhibit broad cytoplasmic expression, suggesting that the exogenous X5P is likely unnecessary in other tissues as well [ , ].

Techniques: