Disease progression and miRNA

People

The PhD student is Hossein Farahani with Jens Lagergren (CIAM/KTH) as the advisor. In addition to the adviser, the reference group consists of Lars Arvestad (SciLIfeLab) and Hugh Salter (AstraZeneca).

Financing

The project is fully funded by CIAM.

Status

Hossein Farahani is expected to finish his PhD in the fall of 2012 with a thesis entitled ''Disease progression and miRNA" (tentative). The project has this far been successful in reached the goals posed at the initial phase. The value of the relation to AstraZeneca established during the project is however very uncertain due to the recently announced closure of the AstraZeneca facility in Södertälje, the unit that was focusing on neurodegenerative diseases.

Background

Modern high-throughput (HTP) experimental biological delivers gigantic datasets at an increasing speed, e.g., genomic, transcriptomic,and proteomic data. AstraZeneca has a special interest in diseases of the brain, e.g., Alzheimers and other neurodegenerative diseases. For the brain there are additional data  sources such as MRI and PET.

Within neurology, for example for Alzheimers and Multiple Sclerosis, and in psychiatry (for example schizophrenia) there is a major need to identify biomarkers that aid diagnosis and prognosis. The input to biomarker identification is multivariate data, including whole genome association studies (genetics), transcriptional profiles of peripheral signals (transcriptomics), protein and metabolite profiles in plasma and CSF (proteomics and metabonomics) and functional data such as PET and MRI. These data forms have a  significant problem in that a) the number of tested variables is extremely large, easily exceeding 1M variables whilst b) there are no established models that can account for all sources of variation in the datasets whilst allowing cross-platform analyses to be carried out.

Multivariate modelling of this problem is likely to be a key to unravelling the underlying biology, since many diseases of interest are widely believed to depend on many small effects combine to cause an effect. The current drive to establish individual genome re-sequencing (which will complicate the problem by at least an  order of magnitude) demands methods that satisfactorily handle the data with robust models, not least as additional variation types (copy numbers, epigenetics etc) need also to be accounted for.

This project is concerned with development of mathematical modelsand methods that facilitate exploitation of HTP data in identification of genes involved in brain diseases.

Goals

There are three goals: (i) development of probabilistic models and analysis algorithms for mutations and regulation of genes, (ii) integration of newly develop methods with publicly available methods, and (iii) application of these analysis methods to in-house AstraZeneca data sets relevant to Alzeimers and other brain related diseases.