Functional metagenomics: AI-based analysis of complex microbial interactions
Project Details
Program
BioEngineering
Field of Study
Bioinformatics
Division
Biological and Environmental Sciences and Engineering
Faculty Lab Link
Center Affiliation
Computational Bioscience Research Center
Project Description
The amount of available protein sequence data is rapidly increasing, for example through applications of sequencing technologies to metagenomics. To understand biological phenomena on a molecular scale, it is crucial to determine the functions of proteins as well as their interactions. Experimental identification of protein functions will not scale to the current and rapidly increasing amount of available protein sequences. Function prediction methods using machine learning may be used to determine protein functions from their sequence. However, proteins rarely function alone but rely on other proteins to perform their function through direct and indirect interactions. The aim of the project is to computationally characterize the functions and interactions of proteins in metagenomes through the development and application of novel AI methods.
About the Researcher
Robert Hoehndorf
Associate Professor, Computer Science
Affiliations
Education Profile
- Ph.D., University of Leipzig, 2009
- M.Sc., University of Leipzig, 2005
Research Interests
a€‹Professor Hoehndorf is interested in artificial intelligence, knowledge representation, biomedical informatics, ontology.Selected Publications
- Hoehndorf, R., Queralt-Rosinach, N., a€œData science and symbolic AI: Synergies, challenges and opportunitiesa€. In: Data Science.
- Boudellioua, I., Mahamad Razali, R. B., Kulmanov, M., Hashish, Y., Bajic, V. B., Goncalves- Serra, E., Schoenmakers, N., Gkoutos, G. V., Schofield, P. N., Hoehndorf, R., a€œSemantic prioritization of novel causative genomic variantsa€. In: PLOS Computational Biology 13.4 (Apr. 2017), pp. 1a-21.
- Hoehndorf, R., Schofield, P. N., Gkoutos, G. V., a€œAnalysis of the human diseasome using phenotype similarity between common, genetic, and infectious diseasesa€. In: Scientific Reports 5 (June 2015), p. 10888.
- Robert Hoehndorf, Tanya Hiebert, Nigel W. Hardy, Paul N. Schofield, Georgios V. Gkoutos, and Michel Dumontier. ""Mouse model phenotypes provide information about human drug targets"". In: Bioinformatics (Oct. 2013).
- Robert Hoehndorf, Paul N. Schofield, and Georgios V. Gkoutos. ""PhenomeNET: a whole-phenome approach to disease gene discovery"". In: Nucleic Acids Research 39.18 (July 2011), e119.
Desired Project Deliverables
Month 1: identification of AI methods, characterization of metagenomics dataset, technical presentation
Month 2: preparation and preprocessing of metagenomics data (QC, assembly)
Month 3: implementation of AI method and data analysis, evaluation
Month 4: combination of AI methods: protein functions and interactions between proteins
Month 5: evaluation results, quantitative characterization
Month 6: write-up
Recommended Student Background
bioinformatics
machine learning
statistics
metagenomics