Tor-Kristian Jenssen

We report a strategy for analysis of data quality in cDNA microarrays based on the repeatability of repeatedly spotted clones. We describe how repeatability can be used to control data quality by developing adaptive filtering criteria for microarray data containing clones spotted in multiple spots. We have applied the method on five publicly available cDNA microarray data sets and one previously unpublished data set from our own laboratory. The results demonstrate the feasibility of the approach as a foundation for data filtering, and indicate a high degree of variation in data quality, both across the data sets and between arrays within data sets.

Abstract Summary: FigSearch is a prototype text-mining and classification system for figures from any corpus of full-text biological papers. The system allows users to search for figures that contain genes of interest and illustrate protein interactions. The retrieved figures are ranked by a score representing the likelihood to be of a certain type, in this case, schematic illustrations of protein interactions and signaling events. The system contains a Web interface for search, a module for classification of figures based on vector representations of figure legends and a module for indexing gene names. In a preliminary validation, the FigSearch system showed satisfactory performance according to domain experts in providing the most relevant graphical representations. This strategy may be easily extended to other figure types. Moreover, as more full-text data become available, such a system will find increased usefulness in identifying and presenting compressed biological knowledge. Availability: A searchable Web interface, FigSearch, is accessible via http://pubgeneserver.uio.no/figsearch/ for all figures from the available corpus.

With the advent of the cDNA microarray and oligonucleotide array technologies it has become possible to study a large number of genes in a single experiment. While experiments with thousands of genes are routinely performed, searching for literature about several genes by traditional methods is time consuming and error-prone. In addition to the inherent limitations of free text search, use of the conventional Boolean operators often result in either none (when AND'ing terms) or far too many (when OR'ing terms) hits. We have created a two-step procedure as an approach to meeting the challenge of multi-gene queries. Our results so far shows that the returned sets of articles scores high on relevance.