Machine learning prediction of cyanobacterial toxin (microcystin) toxicodynamics in humans

Main Article Content

Stefan Altaner
Sabrina Jaeger
Regina Fotler
Ivan Zemskov
Valentin Wittmann
Falk Schreiber
Daniel R. Dietrich

Abstract

Microcystins represent a family of cyclic peptides with approx. 250 congeners presumed to be harmful to human health due to their ability to inhibit ser/thr-proteinphosphatases (PPP), albeit all hazard and risk assessments are based on data of one MC-congener (MC-LR) only. Microcystin congener structural diversity is a challenge for the risk assessment of these toxins, especially as several different PPPs have to be included in the risk assessment. The inhibition of PPP1, PPP2A and PPP5 by 18 structurally different microcystins was determined and demonstrated microcystin congener-dependent inhibition activity and a lower susceptibility of PPP5 to inhibition than PPP1 and PPP2A. The data were employed to train a machine learning algorithm that allows prediction of PPP inhibition (toxicity) based on 2D chemical structure of micro­cystins. IC50 values were classified into three toxicity classes, and three machine learning models were used to predict the toxicity class, resulting in 80-90% correct predictions.

Article Details

How to Cite
Altaner, S., Jaeger, S., Fotler, R., Zemskov, I., Wittmann, V., Schreiber, F. and Dietrich, D. R. (2020) “Machine learning prediction of cyanobacterial toxin (microcystin) toxicodynamics in humans”, ALTEX - Alternatives to animal experimentation, 37(1), pp. 24-36. doi: 10.14573/altex.1904031.
Section
Articles

Most read articles by the same author(s)

1 2 > >>