Meeting report.

I In n s si il li ic co o m me ee et ts s i in n v vi iv vo o Technological developments have had a profound impact on biology during the past decade, spectacularly augmenting our ability to survey and interrogate biological phenomena. In particular, they have increased capacity for data generation by several orders of magnitude and made computation a necessary partner of biology. The sixth meeting in the biennial series of bioinformatics conferences co-sponsored by Georgia Institute of Technology in Atlanta and the Oak Ridge National Laboratory addressed the challenges that this technology-driven avalanche of data pose to bioinformatics-increasing the complexity of long-standing problems and creating new ones. G Ge en no om me e a al li ig gn nm me en nt t a an nd d g ge en ne e p pr re ed di ic ct ti io on n Sequence alignment is unquestionably one of the 'founding problems' in bioinformatics. The availability of sequenced genomes of many species has highlighted the need for methods of making reliable multiple alignments of whole genomes. The alignment of entire genome sequences is much harder to achieve than the alignment of amino-acid sequences of individual proteins, because of the much longer sequences involved (ranging from megabases to tens of megabases), complex evolutionary relationships among the genomes (such as duplications, deletions and translocations) and heterogeneous mutation rates along the sequence. Different methods often produce discrepant alignments with the same set of genomic sequences, and Martin Tompa has attempted to navigate through this complexity. Instead of proposing yet another method for multiple sequence alignment, he presented an approach to evaluating the quality of a given multiple alignment. This is a seemingly more modest goal; he was, however, able to identify high-quality and reliable regions in the multiple alignment, which is very important because downstream comparative genome analysis is compromised by incorrect alignments. Tompa presented data showing that about 10% of the positions in multiple alignments of the human genome with other vertebrate genomes-a widely used technique in comparative genomic studies-are likely to be incorrect. Gene prediction in genomic sequences presents similar problems. Current methods for predicting the exonic structures of protein-coding genes from genomic sequences are generally based on computational models that capture our understanding of the way proteins are encoded in genomes. However, recent surveys of the transcriptional activity of the human genome, …