iMPT-FRAKEL: A Simple Multi-label Web-server that Only Uses Fingerprints to Identify which Metabolic Pathway Types Compounds can Participate In

Abstract

Background: Metabolic pathway is one of the most basic biological pathways in living organisms. It consists of a series of chemical reactions and provides the necessary molecules and energies for organisms. To date, lots of metabolic pathways have been detected. However, there still exist hidden participants (compounds and enzymes) for some metabolic pathways due to the complexity and diversity of metabolic pathways. It is necessary to develop quick, reliable, and non-animal-involved prediction model to recognize metabolic pathways for any compound. Methods: In this study, a multi-label classifier, namely iMPT-FRAKEL, was developed for identifying which metabolic pathway types that compounds can participate in. Compounds and 12 metabolic pathway types were retrieved from KEGG. Each compound was represented by its fingerprints, which was the most widely used form for representing compounds and can be extracted from its SMILES format. A popular multi-label classification scheme, Random k-Labelsets (RAKEL) algorithm, was adopted to build the classifier. Classic machine learning algorithm, Support Vector Machine (SVM) with RBF kernel, was selected as the basic classification algorithm. Ten-fold cross-validation was used to evaluate the performance of the iMPT-FRAKEL. In addition, a web-server version of such classifier was set up, which can be assessed at http://cie.shmtu.edu.cn/impt/index. Results: iMPT-FRAKEL yielded the accuracy of 0.804, exact match of 0.745 and hamming loss of 0.039. Comparison results indicated that such classifier was superior to other models, including models with Binary Relevance (BR) or other classification algorithms. Conclusion: The proposed classifier employed limited prior knowledge of compounds but gives satisfying performance for recognizing metabolic pathways of compounds.

iMPTCE-Hnetwork: A Multilabel Classifier for Identifying Metabolic Pathway Types of Chemicals and Enzymes with a Heterogeneous Network

“…RAndom k-labELsets (RAKEL) [ 15 ] is a classic method to build multilabel classifiers, which can be deemed as the generalization of the Label Powerset (LP) algorithm. To date, this method has been applied to build several multilabel classifiers for tackling different biological and medical problems [ 9 , 21 , 30 – 34 ]. For a multilabel problem involving m labels ( m = 11 in this study), let L = { l 1 , l 2 , ⋯, l m } be its label set.…”

“…According to the results of ten-fold cross-validation, some measurements can be computed. Here, we selected three measurements: accuracy, exact match, and hamming loss [ 9 , 21 , 30 , 31 ]. Their definitions are as below where n stands for the total number of samples, m represents the number of labels, L i and L i ′ denote the actual and predicted label set of the i -th sample, Δ indicates the symmetric difference operation, and ∇ is defined as follows: …”

“…Recently, Baranwal et al [ 8 ] presented a powerful multilabel classifier to assign chemicals to multiple pathways, which adopted graph convolutional networks for obtaining the molecular shape features of chemicals. Jia et al [ 9 ] built a multilabel web server, iMPT-FRAKEL, to predict metabolic pathways of chemicals. This web server had wide applications because it only needed the SMILEs strings of chemicals as the input.…”

MLGL-MP: a Multi-Label Graph Learning framework enhanced by pathway interdependence for Metabolic Pathway prediction

“…Enzymes catalyze drug or drug-like compounds into their metabolites, which differ significantly from these compounds themselves ( Zhang and Tang, 2018 ). As a complex biotransformation, a compound metabolic pathway contains a set of interlocking enzymatic reactions ( Jia et al , 2020a ). In drug discovery, it matters to identify what metabolic pathways a compound attends in the stage of lead compound optimization ( Baranwal et al , 2020 ; Cho et al , 2010 ; Sankar et al , 2017 ).…”

“…Last, they train a set of pathway-specific binary SVMs. As an extension of Guo et al (2018) , Jia et al (2020) recently use random forest (RF) as basic classifiers for accommodating more sub-types of pathways. However, these methods ignore the relationship between labels and aggregate the imbalance between positive and negative samples by generating many negative samples.…”