Abstract:
Personalised nutrition is at its early stages but shows the potential of improving
the health of the general population, at a time when diabetes and obesity are
becoming worldwide epidemics. However, it will need to be based on rigorous
scientific research, as well as being accompanied by public policies and ethical
considerations.
Research is making great progress towards the understanding of the impact of
genetics on complex diseases, which involve hundreds, or thousands, of variants,
each having varying effect on the disease. Personalised medicine aims at
harnessing this genetic information to tailor prevention and treatment according
to each individual.
Unfortunately, the links between the genotype and the phenotype are not yet fully
understood. And while the content of publicly available genetic databases is
exponentially growing, they are often using different formats and means of
access, making it difficult to get complete information. Moreover, evaluating the
genetic predisposition of an individual to a disease is not straightforward, and
while Polygenic Risk Score models can help in this regard, they are often only
based on common variants, which might lead to misevaluation of the risk for rare-
variants carriers.
In this thesis will be presented (i) VarGen, an R package to merge information
from different genetic databases, which has the potential to infer new variant-
disease relationships. (ii) a new method to improve Polygenic Risk Score models,
which includes variants obtained from VarGen on top of the common variants
from standard polygenic analyses. (iii) the results of a microRNA differential
expression analysis, aiming at identifying the impact of microRNAs, on the
development of severe Hypoxic-Ischemic Encephalopathy in new-borns.