Ivo Van Walle
Current interests

Are focused on the multiple alignment problem, both of structures and of sequences. The subject of my PhD thesis is the development of algorithms for this purpose which have improved accuracy and, if possible, speed. In this respect, the Align-m program was created to address the difficult problem of aligning highly divergent sequences.

To evaluate the performance of sequence alignment algorithms, I have constructed the SABmark database, which covers the entire known fold space with alignments of highly divergent sequences.

Publications

Van Walle I., Lasters I. & Wyns L. (2005) An assessment of the limits of sequence and structure alignment with Align-m 2
Bioinformatics, submitted

Van Walle I., Lasters I. & Wyns L. (2005) SABmark - a benchmark for sequence alignment that covers the entire known fold space
Bioinformatics, 21:1267-68

Van Walle I., Lasters I. & Wyns L. (2004) Align-m - a new algorithm for multiple alignment of highly divergent sequences.
Bioinformatics 20:1428-35
Abstract

Loris R., Van Walle I., De Greve H., Beeckmans S., Deboeck F., Wyns L. & Boeckaert J. (2004). Structural basis of oligomannose recognition by the Pterocarpus angolensis seed lectin.
J. Mol. Biol. 335:1227-40
Abstract

Van Walle I., Lasters I. & Wyns L. (2003). Consistency matrices - quantified structure alignments for sets of related proteins.
Proteins 51:1-9
Abstract

Multiple threading

In between sequence and structure alignment lies the threading problem, for which I am developing a 'multiple' environment. This will allow you to align one or more sequences onto one or more fold templates using external pairwise threading algorithm(s). Instead of using only a structure as a fold template, the relaxed transitive alignment (RTA) algorithm allows to supplement this with a weight for each residue that reflects its importance to be aligned. In essence, RTA prodices a similarity score for each residue pair between 2 proteins, put together in a consistency matrix (CM), given N-2 other, similar proteins. A CM can be compared to a dot-plot matrix but instead contains values between 0 and 1 instead of either 0 or 1. From these CMs, a score can be assigned to the residues of a single protein, which reflects the importance of the spatial position they occupy. RTA is currently not separately available for download, but feel free to contact me if you wish to use it.

The pictures below are examples of resp. a CM and a fold template: