The use of gene clusters to infer functional coupling

Abstract

Previously, we presented evidence that it is possible to predict functional coupling between genes based on conservation of gene clusters between genomes. With the rapid increase in the availability of prokaryotic sequence data, it has become possible to verify and apply the technique. In this paper, we extend our characterization of the parameters that determine the utility of the approach, and we generalize the approach in a way that supports detection of common classes of functionally coupled genes (e.g., transport and signal transduction clusters). Now that the analysis includes over 30 complete or nearly complete genomes, it has become clear that this approach will play a significant role in supporting efforts to assign functionality to the remaining uncharacterized genes in sequenced genomes.Gene clusters are known to be prominent features of bacterial chromosomes. Demerec and Hartman (1) postulated in 1959 that ''regardless of how the gene clusters originated, natural selection must act to prevent their separation'' and the ''mere existence of such arrangements shows that they must be beneficial, conferring an evolutionary advantage on individuals and populations which exhibit them.'' One of the most striking features of prokaryotic gene clusters is that typically they are composed of functionally related genes. For the past 40 years, there has been vigorous, ongoing discussion on the functional significance of gene arrangement on the chromosome, as well as the origin and mechanisms of maintenance of gene clusters (see, for example, refs. 2-5).Here, we present a method that uses conserved gene clusters from a large number of genomes to predict functional coupling between genes in those genomes. This article further develops the approach that we previously reported (6) and uses this method to reconstruct several major metabolic and functional subsystems. MethodologyThe data presented below are computed via the WIT system (http:͞͞wit.mcs.anl.gov͞WIT2͞), developed by Overbeek et al. (7) at Argonne National Laboratory. WIT was designed and implemented to support genetic sequence analysis, metabolic reconstructions, and comparative analysis of sequenced genomes; it currently contains data from over 30 genomes, albeit a few of them are incomplete.Our approach to detection of conserved clusters of genes is based on the following definitions: a set of genes occurring on a prokaryotic chromosome will be called a ''run'' if and only if they all occur on the same strand and the gaps between adjacent genes are 300 bp or less. Any pair of genes occurring within a single run is called ''close.'' Given two genes X a and X b from two genomes G a and G b , X a and X b are called a ''bidirectional best hit (BBH)'' if and only if recognizable similarity exists between them (in our case, we required FASTA3 scores lower than 1.0 ϫ 10 Ϫ5 ), there is no gene Z b in G b that is more similar than X b is to X a , and there is no gene Z a in G a that is more similar than X a is to Computation of PCBBHs for 31 complete or ne...

Biosynthesis of Isoprenoids via Mevalonate in Archaea: The Lost Pathway

“…Analysis of gene colinearity in completely sequenced microbial genomes has shown that the longrange conservation of the order of orthologous genes on a chromosome is observed only between closely related species (Tatusov et al 1996;Overbeek et al 1999). At the other extreme, several small sets of orthologous genes are arrayed in the same order in most bacteria and archaea, and analysis of the products encoded by these universally conserved operons suggests that they code for the subunits of stoichiometric protein complexes Dandekar et al 1998).…”

“…Comparative protein sequence analysis, with the attention to both close and more distant similarities, remains the principal method of function prediction and pathway reconstruction (Tatusov et al 1996;Koonin et al 1997;Bork et al 1998). In the attempt to break through the "similarity barrier", complementary approaches have been proposed, including identification of genes that are physically close, and presumably coordinately regulated, in multiple genomes (Overbeek et al 1999), and the use of phyletic profiles, i.e., the occurrence in some genomes of sets of genes that are missing in others (Tatusov et al 1997;Pellegrini et al 1999). We are interested in applying all available methods to reconstruct pathways in poorly characterized species, in order to develop a highly automated strategy facilitating this process and to get an insight in general and specific trends in the evolution of biochemical pathways.…”

Individual Gene Cluster Statistics in Noisy Maps

“…The problem of identifying true homologs has been much debated and numerous solutions have been proposed [33,43,[48][49][50]. The first step is typically sequence comparison.…”

“…The intuitive notion that gene clusters share similar but not conserved gene content has been translated into a number of different formal models for finding and testing clusters [6,29,[33][34][35][36][37][38][39][40][41][42][43][44][45], yet most of these are not suitable for individual clusters. Cluster statistics depend on the size of the search space.…”

Fungal Genomic Annotation

“…This relationship is calculated with all the proteins in the genome. The logic in performing this is as follows: in two genomes A and B containing genes Xa and Xb, respectively, Xa and Xb are potential orthologs if there is no better alignment to Xa from genome B than Xb, and there is no better alignment to Xb from genome A than Xa (Lee et al 2002, Overbeek et al 1999. COGs (Tatusov et al 2001) extends this approach by requiring that orthologs be from three genomes ("triangles" of proteins termed BeTs) to be considered orthogous, thus ensuring that the gene has persisted through time.…”