Klinisk institutt 1

Midtveisevaluering - Fiona Dick

Midtveisevaluering for ph.d.-graden ved Universitetet i Bergen for kandidat Fiona Dick

Fiona Dick er tilknyttet Klinisk institutt 1. Veiledere er Charalampos Tzoulis og Gonzalo Sanchez Nido.


Differential transcript usage in aging and neurodegeneraiton


Parkinson’s disease (PD) is the second most prevalent neurodegenerative disorder, affecting∼1.8 % of the population above 65 years. PD is a complex disorder caused by a combination ofgenetic and environmental factors, but the molecular mechanisms underlying its etiology remainlargely unaccounted for. Previous gene expression studies carried out in brain tissue of PDpatients have largely been guided by specific hypotheses and restricted to genes of interest.Transcriptome-wide studies, on the other hand, enable us to study differences in geneexpression as well as differential transcript usage (DTU) in a hypothesis free framework.Limitations, however, remain when analysing bulk tissue. Possible sources of bias intranscriptome analyses of bulk tissue are differences in RNA quality and cellular composition.These can arise from both biological and technical variability with the latter being introducedduring sample preparation. We analyzed RNA sequencing data from brain tissue of N=27 PDindividuals and N=22 healthy controls. We accounted for confounding sources of variance byincluding both the estimates for cellularity and a proxy for RNA quality as covariates in ourdifferential expression models. In addition to changes in gene expression, we aimed atcharacterizing the isoform usage landscape associated with PD in our dataset. Our resultsreveale d that DTU does occur in PD. Interestingly, genes that showed an altered transcriptusage profile were generally not identified as differentially expressed at the gene-level.Furthermore, we found a disagreement in the direction of change between gene- andtranscript-level analyses for the majority of genes that were classified as differentiallyexpressed. These results highlight the necessity of DTU as a complementary analysis totraditional differential expression. The l ack of transcript-level information can otherwise lead toincorrect assumptions on the magnitude and direction of effect size and potentially result ininaccurate predictions of the functional consequences.

In order to dissociate normal aging isoform patterns from PD-associated DTU, we are currentlystudying an additional cohort of healthy individuals of ages ranging from under the age of 1 upto 88 years.