[metagenomics-jclub] Next meeting, 10/19/11, noon PSB271, Yuan Zhang

[Adina Chuang Howe]

Please join us as Yuan Zhang (from Yanni Sun's lab) presents a novel
approach for protein domain classification for metagenomic sequences
to help us who are interested in scalable, accurate sequence
annotations.  We'll be meeting next Wednesday at noon (please feel
free to bring your own lunch) at PSB 271.

Summary:

Metagenomics is a new field of research which makes use of
next-generation sequencing technologies (NST) to sequence microbial
genomes directly recovered from environmental samples. It enhances
people's view of the microbial world. However, there are several
challenges in metagenomics which are yet to be handled. First,  high
throughput sequencing technologies generate huge volumes of data sets
which need efficient algorithms for mapping or assembly. Second,
fragmented short reads make it difficult to do homology search or
assembly. Finally, frameshifts introduced by sequencing errors make
alignment tools such as HMMER generate marginal scores and thus these
tools may miss most hits. In order to address these challenges, we
developed two profile HMM-based protein analysis tools, HMM-FRAME and
MetaDomain. The former one is used to detect and correct sequencing
errors and thus improve performance of profile HMM-based alignment
tools. The latter is able to align very short reads to their native
Pfam domains and evaluate expression levels of Pfam domains
represented by Metagenomic data sets. Experimental results show that
these tools greatly improve performance of protein domain analysis
compared to existing tools. However, there is still space for future
improvement.

Hope to see you there!
Adina