MASAMB XXI, in Vienna

I was pleased to be invited to present some of my work at the MASAMB conference in Vienna recently (11-12^th April 2011). My talk was titled “Quantifying the effect of evolution and genomic alignment on de novo RNA gene prediction”. I have a similarly themed manuscript in preparation, so won’t go into detail here. But if you have any questions about the presentation, then please get in touch by
. The abstract is included below the nice pics.

I enjoyed both the conference and the time I spent in Vienna, where I appreciated the architecture and culture, as well as the beer and the wildlife:

Quantifying the effect of evolution and genomic alignment on de novo RNA gene prediction

James Allen and Simon Whelan

University of Manchester, Faculty of Life Sciences, UK

Non-coding RNA is biologically important and linked to a range of diseases, yet there are potentially many RNA genes of unknown function that do not belong to characterized RNA families. The de novo prediction of RNA genes is difficult because they are heterogeneous, and tend to have fewer restrictions than protein-coding genes. These problems are compounded by the lack of independent datasets for benchmarking predictions. We identify alignments of genes from 9 well-defined RNA families in UCSC genomic data for 32 vertebrate species, totaling 296 gene regions, and use randomization to create appropriate negative-control datasets. We evaluate three popular prediction programs (CMfinder, EvoFold, and RNAz), testing their ability to detect RNA genes and accurately determine gene boundaries. Our results show that RNA genes in genomic alignments have different evolutionary characteristics to structurally-aware RNA alignments, affecting gene prediction, and provide practical suggestions for the development and application of prediction programs.