Accurate, automatic annotation of peptidases with Hotpep-protease

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review


Peptidases are essential for intracellular protein processing, signaling and homeostasis, physiological processes and for digestion of food. Moreover, peptidases are important biotechnological enzymes used in processes such as industrial food processing, leather manufacturing and the washing industry. Identification of peptidases is a crucial step in their characterization but peptidase annotation is not a trivial task due to their large sequence diversity.

In the present study short, conserved sequence profiles were generated for all peptidase families with more than four members in the comprehensive Merops peptidase database. The sequence profiles were combined with the Homology to Peptide Pattern (Hotpep) method for automatic annotation of peptidases. This method is a stand-alone software that annotates protease sequences to Merops family and subgroup and is suitable for large-scale sequence analysis. Compared to the Mammalian Degradome Database Hotpep-protease had an accuracy of 92% and a sensitivity of 96% for annotation of the human degradome. Annotation by commonly used methods (Blast and conserved domains) had an accuracy of 69% and a sensitivity of 78%. For fungal genomes, there were large differences between annotation with Hotpep-protease, Blast- and Hidden Markov Model-based annotation and the Merops annotation, which confirms the difficulty of large-scale peptidase annotation. Manual annotation indicated that Hotpep-protease had a positive prediction rate of 0.90 compared to a positive prediction rate of 0.67 for Blast search. Hence, Hotpep-protease is highly accurate method for fast and accurate annotation of peptidases.
TidsskriftGreen Chemical Engineering
Udgave nummer2
Sider (fra-til)124-130
Antal sider7
StatusUdgivet - 28 nov. 2020

Citer dette