Computational methods for analyzing small RNAs and their interaction partners with large-scale techniques

This thesis describes the computational tools and analyzes developed to characterize
small regulatory RNAs and their interaction partners using large-scale techniques.
Following an introduction into the emerging world of small regulatory RNAs, our methodology
for analyzing small RNAs from deep-sequencing data is described (chapter 2). This methodology
allows the classification of small RNAs obtained by sequencing and provides several
downstream analysis tools such as expression profiling and miRNA sample comparison.
It has been integrated with a miRNA target prediction method
into a web server which allows one to explore tissue-specific miRNA targeting (chapter 3).
In the fourth chapter, an experimental procedure for genome-wide identification
of miRNA targets is outlined. With this procedure, we identified the mRNAs,
that are targeted by the most abundant miRNAs in HEK293 cells. Importantly,
the experimental protocol enabled us to identify the exact location of the
miRNA-mRNA interaction site within the mRNA as well as the precise position
of the mRNA-protein crosslink.
The fifth and sixth chapter describe our studies of murine embryonic stem cells
and oocytes that are devoid of Dicer.
The murine specific miR-290 cluster has been identified as an important regulator
in embryonic stem cells.
The loss of pluripotency in Dicer-/- embryonic stem cells has been linked
to primary and secondary targets of the miR-290 cluster. In contrast,
our analysis
of Dicer-/- oocytes revealed that the miRNA pathway plays only a minor part during
oocyte maturation, and loss of Dicer affects mainly the endo-siRNA pathway.
Finally, we reanalyzed piRNA sequence reads from various species (chapter 6). This
analysis revealed an unexpected 19 nt long processing product which is generated during piRNA biogenesis.