engleski

De novo transcriptome assembly

In this thesis, a de novo transcriptome assembler was implemented based on the overlap-layout-consensus paradigm. It was written in the C++ programming language and was named Ra which is short for RNA assembler. Its overlap phase relies on the enhanced suffix arrays and reproduces exact overlaps between input reads. The layout phase uses several methods for graph simplification which includes trimming and bubble popping. Due to the exact overlap phase there is no need for a consensus phase at this moment but there exists one which is based on the partial order alignment algorithm. Conducted tests have shown that Ra needs improvements to compete with other transcriptome assemblers. Source code is available at https://github.com/rvaser/ra.

RNA; transcriptome assembly; overlap-layout-consensus; suffix array; graphs

nije evidentirano