Center for Omic Sciences
Nutrigenomics and Personalised Nutrition
Integration of omic data for personalized nutrition
Omics data have a great potential for personalized nutrition, and while there already are some projects that leverage them to some extent, it still has a lot of space for exploration.
The idea of this project is to use data from diferent omics sources (metabolomics, metagenomics, proteomics, etc.), from different kinds of samples (urine, fecal matter, blood, etc.), combined from other data from patients/user (like clinical data or eating habits, etc.) to build a predictive model (through machine learning or similar methods) capable of suggesting the best food choices for a patient’s diet in order to imrpove or maintain their health, according to their conditions.
Omics sciences that would provide insight about the status of the patient beyond tools already available, while also only requiring the easiest and least invasive sampling to get from patients (urine and fecal samples), would allow the following information to be incorporated into the model:
• Metabolomics (urine): would enable the tracking of certain biomarkers in the patient’s urine, which would allow for the prediction of health changes or to show a treatment’s effect on the patient’s health over time, or whether a treatment needs to be reapplied, etc.
• Metagenomics (fecal): would enable the construction of a gut microbiome map for the patient. The importance of the gut microbiome and its healthy equilibrium is becoming more clear as it’s being studied more closely; changes in diet or treatments will reflect in changes within the microbiome, and other health changes can be reflected there – or even caused by changes in gut microbiome. A very useful marker to include in the model.
• Metaproteomics (fecal): a metaproteomics approach can shed light on which kind of enzymes are currently produced by the patient’s microbiome, which can change considerably even without changes in microbiome population, and which can also be a marker for disease or health tracking.
• Proteomics (urine): would make it possible to track the presence of certain proteins in urine in case of dealing with patients with proteinuria
• Genomics: from samples containing DNA, a sequentiation of certain genes or regions of interest, known to be linked to certain diseases, could be possible.
• Metabolomics: with blood samples, blood concentrations of many different metabolites could be used for the model
• Transcriptomics: with blood or other tissue samples, the levels of expression of certain genes could be used for the model.
BSc. or MSc. in Computer Science or similar; OR BSc. or MSc. in Life Sciences
Proficiency in at least 1 general-purpose programming language (C/C++, Java, Perl, Python, etc.)
Ability to research independently and a high level of theoretical understanding.
Ability to learn and use new tools and workflows
Desirable, but not required:
Proficiency in Python
Familiarity with pandas, scipy, scikit-learn, theano and related modules
37.5 hours a week
|This project has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 713679|