MOTIVATION: Identifying regulatory modules is an important task in the exploratory analysis of gene expression time series data. Clustering algorithms are often used for this purpose. However, gene regulatory events may induce complex temporal features in a gene expression profile, including time delays, inversions and transient correlations, which are not well accounted for by current clustering methods. As the cost of microarray experiments continues to fall, the temporal resolution of time course studies is increasing. This has led to a need to take account of detailed temporal features of this kind. Thus, while standard clustering methods are both widely used and much studied, their shared shortcomings with respect to such temporal features motivates the work presented here. RESULTS: Here, we introduce a temporal clustering approach for high-dimensional gene expression data which takes account of time delays, inversions and transient correlations. We do so by exploiting a recently introduced, message-passing-based algorithm called Affinity Propagation (AP). We take account of temporal features of interest following an approximate but efficient dynamic programming approach due to Qian et al. The resulting approach is demonstrably effective in its ability to discern non-obvious temporal features, yet efficient and robust enough for routine use as an exploratory tool. We show results on validated transcription factor-target pairs in yeast and on gene expression data from a study of Arabidopsis thaliana under pathogen infection. The latter reveals a number of biologically striking findings. AVAILABILITY: Matlab code for our method is available at http://www.wsbc.warwick.ac.uk/stevenkiddle/tcap.html.

MOTIVATION: We compare phylogenetic approaches for inferring functional gene links. The approaches detect independent instances of the correlated gain and loss of pairs of genes from species' genomes. We investigate the effect on results of basing evidence of correlations on two phylogenetic approaches, Dollo parsminony and maximum likelihood (ML). We further examine the effect of constraining the ML model by fixing the rate of gene gain at a low value, rather than estimating it from the data. RESULTS: We detect correlated evolution among a test set of pairs of yeast (Saccharomyces cerevisiae) genes, with a case study of 21 eukaryotic genomes and test data derived from known yeast protein complexes. If the rate at which genes are gained is constrained to be low, ML achieves by far the best results at detecting known functional links. The model then has fewer parameters but it is more realistic by preventing genes from being gained more than once. AVAILABILITY: BayesTraits by M. Pagel and A. Meade, and a script to configure and repeatedly launch it by D. Barker and M. Pagel, are available at http://www.evolution.reading.ac.uk

Microsatellites are repetitive regions of DNA which account for a highly significant amount of eukaryotic genomes. In this study we use a vast amount of microsatellite data from the human genome, and use an evolutionary algorithm to search for DNA patterns which might yield clues about their generation and perhaps their inhibition too. We find that the evolutionary algorithm uncovers several significant patterns which indicate preferences for microsatellites to appear nearby, and other patterns which seem to inhibit microsatellite production. The results show that both attractors and inhibitors to microsatellite production seem to exist, and are motif-dependent. We feel this represents a nice example of using evolutionary algorithms to find previously unimagined fuel for molecular biology hypothesis making.

DNA annotation is an important problem in molecular biology. The task in annotation is to determine functional roles of DNA regions. We are interested in core promoter regions in eukaryotes; these are approximately 50 bp in length, and are believed to act as the critical factor in attracting the transcription initiation complex to the DNA at that site, so promoting the transcription of a gene immediately downstream. The pattern of nucleotides which constitute a core promoter region is very complex, resulting in relatively poor performance from classification methods so far in terms of distinguishing true core promoter regions from other regions of DNA. The methods described both use evolutionary algorithms to optimise simple models of core promoter regions. They seem to perform comparably to the best of many published methods so far, but has the advantage of simplicity and transparency. That is, we are able to evolve easily understandable models of the core promoter region which perform comparably with the best of existing techniques. There seem to be several ways in which the scheme can be further developed.

Phylogenetics is the study of evolutionary relationships between and among species or individuals, and involves the construction and analysis of evolutionary trees. In particular, certain questions frequently arise which can be examined by analysing a tree in terms of maximum likelihood with respect to a particular characteristic or set of characteristics. An important concept in these respects is the 'transition rate' between two states of a character over evolutionary time. Various analyses based on given phylogenetic trees involve a search for a set of transition rates which maximise the likelihood of the given tree. We investigate the preliminary use of an evolutionary algorithm for this search problem, and find that it seems significantly superior to the existing method (a conjugate-gradient search) used in the evolutionary biology community for this purpose. Our test case is drawn from the field of evolutionary linguistics, emphasising the broad applicability of this search problem.

Mobile genetic elements, portions of DNA which are able to copy themselves elsewhere into the genome, have played a substantial role during evolution. One of the most prominent such elements is the LINE-1 retrotransposon, a section of DNA around 6,000 nucleotide bases in length, which is thought to account for about 15% of the current state of the human genome. The mechanism of LINE-1 insertion is rather poorly understood. However, achieving a good understanding of this process is fundamental to understanding natural evolution at the molecular level. We describe a first approach to use evolutionary computation to explore models for the LINE-1 insertion process. A range of findings from standard genome studies are able to suggest the basic parameters and structure of an insertion model. We use an evolutionary algorithm to explore a space of such models.