Leveraging heterogeneous network embedding for metabolic pathway prediction

Abstract

Motivation Metabolic pathway reconstruction from genomic sequence information is a key step in predicting regulatory and functional potential of cells at the individual, population and community levels of organization. Although the most common methods for metabolic pathway reconstruction are gene-centric e.g. mapping annotated proteins onto known pathways using a reference database, pathway-centric methods based on heuristics or machine learning to infer pathway presence provide a powerful engine for hypothesis generation in biological systems. Such methods rely on rule sets or rich feature information that may not be known or readily accessible. Results Here, we present pathway2vec, a software package consisting of six representational learning based modules used to automatically generate features for pathway inference. Specifically, we build a three layered network composed of compounds, enzymes, and pathways, where nodes within a layer manifest inter-interactions and nodes between layers manifest betweenness interactions. This layered architecture captures relevant relationships used to learn a neural embedding-based low-dimensional space of metabolic features. We benchmark pathway2vec performance based on node-clustering, embedding visualization and pathway prediction using MetaCyc as a trusted source. In the pathway prediction task, results indicate that it is possible to leverage embeddings to improve pathway prediction outcomes. Availability and implementation The software package, and installation instructions are published on github.com/pathway2vec Supplementary information Supplementary data are available at Bioinformatics online.

Multi-label classification with XGBoost for metabolic pathway prediction

“…The single organism golden dataset consists of six Tier 1 PGDBs from BioCyc which are EcoCyc(v21) [ 28 ], HumanCyc(v19.5) [ 4 ], AraCyc(v18.5) [ 29 ], YeastCyc(v19.5), LeishCyc(v19.5) [ 30 ] and TrypanoCyc(v18.5) [ 31 ] which were used in previous benchmarks [ 9 , 11 , 17 ]. BioCyc is a PGDB Web portal that contains thousands of PGDBs and divides PGDBs into tiers based on the manual curation involved [ 32 ].…”

“…Only metabolic pathways are predicted instead of other types of biochemical pathways such as signaling pathways. Metabolic pathway prediction in the literature commonly refers to predicting either the metabolic pathways that a molecule is associated with [ 5 – 7 ] or the metabolic pathways occurring in an organism based on its annotated genome [ 1 , 8 , 9 ]. This work will focus on the latter and assumes that the reactome is already inferred and provided.…”

“…A limitation of previous machine learning-based metabolic pathway prediction methods was that the feature engineering task involving designing and testing features was a time consuming task. As a response to this limitation, representational learning approaches [ 16 ] such as pathway2vec [ 9 ] and triUMPF [ 17 ] were introduced to generate features to be used for prediction. While the research direction and results from the two studies are promising, they shared similar problems with mlLGPR in their evaluation methodology for single organism genomes.…”

Graph embedding on mass spectrometry- and sequencing-based biomedical data

“…Metabolic pathways, in which each product (i.e., node) is solely dependent on its precursor (i.e., a previous node in the pathway), can be defined as directed. However, most metabolic pathways are represented as undirected graphs, since their chemical reactions are reversible and regulated by feedback loops, where downstream products influence the formation of their upstream precursors (e.g., in glycolysis) [ 33 , 34 ].…”

“…While Deepwalk and Node2vec provided a solution to tasks such as link prediction, node classification, node clustering (community detection), and visualization, two random-walk algorithms, Netpro2vec [ 92 ] and Pathway2vec [ 33 ], were proposed to better analyze biomedical datasets.…”