Understanding the mechanisms of client protein interaction with Hsp70 chaperones is essential to analyze the complex dynamics in the context of normal or dysregulated metabolism. Because Hsp70 can bind millions of proteins, including key molecules involved in processes of stemness, tumorigenesis and survival, in silico prediction of Hsp70 interactions has great value in validating possible new clients. Currently, two algorithms are available to predict binding to DnaK—the bacterial Hsp70—but both are based on amino acid sequence and energy calculations of qualitative information—binders and non-binders.


We introduce a new algorithm to identify Hsp70 binding sequences in proteins—ChaperISM—a position-independent scoring matrix trained on either qualitative or quantitative chemiluminescence data previously published, which were obtained from the interaction between DnaK and different ligands. Both versions of ChaperISM, qualitative or quantitative, resulted in an improved performance in comparison to other state-of-the-art chaperone binding predictors.

Availability and implementation

ChaperISM is implemented in Python version 3. The source code of ChaperISM is freely available for download at

Supplementary information

Supplementary data are available at Bioinformatics online.

This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (