Agnieszka Gierlik

Total number of coding open reading frames in the yeast genome

At the end of 1996 we approximated the total number of protein coding ORFs in the Saccharomyces cerevisiae genome, based on their properties, as 4700–4800. The number is much smaller than the 5800 which is widely accepted. According to our calculations, there remain about 200–300 orphans—ORFs without known function or homology to already discovered genes, which is only about 5% of the total number of genes. Our results would be questionable if the analysed set of known genes was not a statistically representative sample of the whole set of protein coding genes in the S. cerevisiae genome. Therefore, we repeated our estimation using recently updated databases. In the course of the last 18 months, previously unknown functions of about 500 genes have been found. We have used these to check our method, former results and conclusions. Our previous estimation of the total number of coding ORFs was confirmed. Copyright

Evolution Rates of Genes on Leading and Lagging DNA Strands

Abstract. One of the main causes of bacterial chromosome asymmetry is replication-associated mutational pressure. Different rates of nucleotide substitution accumulation on leading and lagging strands implicate qualitative and quantitative differences in the accumulation of mutations in protein coding sequences lying on different DNA strands. We show that the divergence rate of orthologs situated on leading strands is lower than the divergence rate of those situated on lagging strands. The ratio of the mutation accumulation rate for sequences lying on lagging strands to that of sequences lying on leading strands is rather stable and time-independent. The divergence rate of sequences which changed their positions, with respect to the direction of replication fork movement, is not stable—sequences which have recently changed their positions are the most prone to mutation accumulation. This effect may influence estimations of evolutionary distances between species and the topology of phylogenetic trees.

Mechanisms generating long-range correlation in nucleotide composition of the Borrelia burgdorferi genome

We have analysed protein coding and intergenic sequences in the Borrelia burgdorferi (the Lyme disease bacterium) genome using different kinds of DNA walks. Genes occupying the leading strand of DNA have significantly different nucleotide composition from genes occupying the lagging strand. Nucleotide compositional bias of the two DNA strands reflects the aminoacid composition of proteins. 96% of genes coding for ribosomal proteins lie on the leading DNA strand, which suggests that the positions of these as well as other genes are non-random. In the B. burgdorferi genome, the asymmetry in intergenic DNA sequences is lower than the asymmetry in the third positions in codons. All these characters of the B. burgdorferi genome suggest that both replication-associated mutational pressure and recombination mechanisms have established the specific structure of the genome and now any recombination leading to inversion of a gene in respect to the direction of replication is forbidden. This property of the genome allows us to assume that it is in a steady state, which enables us to fix some parameters for simulations of DNA evolution.

The Differential Killing of Genes by Inversions in Prokaryotic Genomes

We have elaborated a method which has allowed us to estimate the direction of translocation of orthologs which have changed, during the phylogeny, their positions on chromosome in respect to the leading or lagging role of DNA strands. We have shown that the relative number of translocations which have switched positions of genes from the leading to the lagging DNA strand is lower than the number of translocations which have transferred genes from the lagging strand to the leading strand of prokaryotic genomes. This paradox could be explained by assuming that the stronger mutation pressure and selection after inversion preferentially eliminate genes transferred from the leading to the lagging DNA strand.

Effect of replication on the third base of codons

We have analyzed third position in codons and have observed strong long-range correlations along DNA sequence. We have shown that the correlations are caused mostly by asymmetric replication. In the analysis, we have used a DNA walk (spider analysis Cebrat et al., Microbial Comparative Genomics 2(4)(1997) 259–268) in two-dimensional space [A–T,G–C]. The particular case of the E.coli sequence has been studied in detail.

Optimization of gene sequences under constant mutational pressure and selection

We have analyzed the influence of constant mutational pressure and selection on the nucleotide composition of DNA sequences of various size, which were represented by the genes of the Borrelia burgdorferi genome. With the help of MC simulations we have found that longer DNA sequences accumulate much less base substitutions per sequence length than short sequences. This leads us to the conclusion that the accuracy of replication may determine the size of genome.

Some Hints On Open Reading Frame Statistics — How Orf Length Depends On Selection

Coding sequences of DNA generate Open Reading Frames (ORFs) inside them with much higher frequency than random DNA sequences do, especially in the antisense strand. This is a specific feature of the genetic code. Since coding sequences are selected for their length, the generated ORFs are indirect results of this selection and their length is also influenced by selection. That is why ORFs found in any genome, even much longer ones than those spontaneously generated in random DNA sequences, should be considered as two different sets of ORFs: The first one coding for proteins, the second one generated by the coding ORFs. Even intergenic sequences possess greater capacity for generating ORFs than random DNA sequences of the same nucleotide composition, which seems to be a premise that intergenic sequences were generated from coding sequences by recombinational mechanisms.