From: Park, Yonil (NIH/NLM/NCBI) [E]
Sent: Thursday, February 12, 2009 5:13 PM
To: NLM/NCBI List ncbi-seminar
Subject: CBB seminar, Feb. 17, 11AM 

Time: Tuesday, February 17, 2009, 11AM
Location: NCBI Library, B2 floor, Building 38A
Speaker: Yonil Park

Title: The Problem of Spurious Flanks in Sequence Alignment

Pairwise sequence alignment is a ubiquitous tool for learning about the
evolution and function of DNA, RNA, and protein sequences.  It is
therefore essential to avoid spurious alignments, and statistical
techniques have been developed to eliminate alignments that wholly arise
by chance.  However, it is not known how to avoid alignments that are
partly spurious, and it is not even known how often partly spurious
alignments occur. 

In this talk, I show that some commonly used alignment scoring schemes
are severely prone to over-extension, often by hundreds of bp/aa. For
example, spurious extensions likely comprise > 18% of the human-fugu
genome alignment in the UCSC genome database.  I provide guidance on
choosing appropriate scoring schemes, by showing the tradeoff between
over-alignment and under-alignment.  Finally, I provide a simple
over-alignment p-value. The over-alignment p-value can identify spurious
alignment flanks and help us to avoid making false inferences about
evolution and function from spurious extensions of alignments.

FLANK, software for over-alignment parameters estimation, is available
at
http://www.ncbi.nlm.nih.gov/CBBresearch/Spouge/html/index/software.html#
FLANK


----------
Yonil Park 
NCBI/NLM/NIH
Bldg 38A, Room 6N611N
8600 Rockville Pike
Bethesda, MD 20894
Voice: 301-402-1438
Fax:   301-480-2288