Our algorithm divides genes and properties (usually one time point or condition in an experiment such a DNA microarray) into Bi- clusters ("bics") composed of genes and properties such that the genes behave similarly in that group of conditions. Bics may overlap (in genes, conditions, or both), and a gene may be present in several bics, as it may be similar to one group of genes under conditions A-G and to a different set of genes under conditions H-M. Properties may be anything obtainable from genome-wide experiments: level of RNA expression, does or does not interact physically, does or does not interact genetically, is it or is it not resistant/sensitive to a certain drug, etc.The Universe of Bics can be explored using three different views:
ESR: Many treatments cause a standard response ("Environmental Stress Response", Mol Biol Cell. 2000 Dec;11(12):4241-57. Genomic expression programs in the response of yeast cells to environmental changes. Gasch AP, Spellman PT, Kao CM, Carmel-Harel O, Eisen MB, Storz G, Botstein D, Brown PO.), in which a large number of genes is shut-off (usually those involved in growth) and a large number of genes become highly expressed (chaperons, etc.). This stress response is elicited by many, many types of treatment and is highly visible in the Bic Universe (see Integrative analysis of genome-wide experiments in the context of a large high-throughput data compendium. Tanay A, Steinfeld I, Kupiec M, Shamir R. Mol Syst Biol. 2005;1:2005.0002. Epub 2005 Mar 29.). Sometimes people would like to filter-out this response, so we offer the possibility of choosing the fraction of bics labeled ESR. If you choose 0, you eliminate them all. We recommend SOME level of ESR bics (e.g., 0.1 or 0.2).