������

General workflow of metabolite identification

Metabolites are small molecules that can provide information about the state of the cells. Metabolite identification is an important problem in metabolomics. A contest on the identification of small molecules from mass spectrometry data was organized in the spring 2016. The objective of the CASMI-contest was to provide a common framework for evaluating different metabolite identification methods.

For the first time this year, a challenge for automatic methods was proposed. The goal of this challenge was to determine the correct molecular structure among a list of potential candidates for approximately 200 molecules. 

The category of the best automatic approach, in the case where additional information can only be used to train the prediction model, was won by researchers from Aalto University and the University of Jena in Germany. They predicted a better score than the other teams for 86 of the molecules. The Input Output Kernel Regression (IOKR) approach used was recently published in the Bioinformatics journal. Team members of the winning team were Céline Brouard, Huibin Shen, Kai Dührkop, Sebastian Böcker and Juho Rousu.

Team Duhrkop members were Kai Dührkop, Huibin Shen, Sebastian Böcker and Juho Rousu and they used CSI:FingerID approach, coming second in the category.

More information:

“Fast metabolite identification with Input Output Kernel Regression”,