The Distribution Of Transcription Factor Binding Sites In Upstream Regions Of Yeast Genes

Gene transcription regulation is one of the most important fields in the biology research. Transcription factors recognize degenerate families of short DNA sequences (5-25 base pair) to regulate expression of the target gene. A variety experiments have manifested that transcription factors binding sites mainly exist in the upstream region of the transcription start site (TSS). In order to acquire the distribution of the transcription factors binding sites in the upstream region of TSS, this article studies the transcription factor structure, the function character of target gene, the sites conservation by bioinformatics method.First, we investigate the binding sites distributions of transcription factors in basic domain super class and beta-Scaffold super class in per 100 bp in the 2000 bp upstream region from TSS. After that, the genes which have different function regulated by transcription factors were clustered, based on four levels, such as biological process, molecular function, cellular component, and major function. Then we analyzed the different distribution of binding sites in different groups. Finally, we calculate the sites conservation indexes Mi and PWM of different regions.The distribution of binding sites for transcription factors reveals that they are not uniformly distribution over the upstream region but instead shows a sharply peaked distribution. The majority (61%) lie in 100~500 bp upstream of the gene. Our categorization strongly suggests that factors of similar structure (in one family) have similar binding sites distribution, while other transcription factors binding sites distribution are significant different(P<0.01). Except for a few transcription factors, the binding sites of majority transcription factors (19/22) are not significant different (P>0.05) in genes of different classes in function. These show that structure of transcription factors makes a great impact on the distribution of binding sites of the transcription factors, while the function of the genes has a little effect on that, the biological processes of the gene and the cellular component where the gene in are irrelevant with the distribution of the binding sites.Based on the differences of nucleotide probability in every position of transcription factor binding sites, the sites conservation indexes Mi are calculated in the binding sites of four transcription factors which regulate most genes. The results show: (1) the Mi in every position of transcription factor binding sites are quite different, which change irregularly, and only some strongest conservation sites change mildly; (2) the conservation of consensus sequence changes very complicated in different regions, while in some regions have high score, but no obvious rule; (3) the Mi distributions of the four transcription factors are significant difference. Therefore, the relationship between the conservation of the consensus model and the regions they binding is weak, and probably relative to the transcription factor DNA binding domain structure.Previous studies mainly focus on the existence of class-specific features in the DNA sequences bound by each class of transcription factors, this article reveals that the binding sites of transcription factors often have family-specific feature in the distribution in the upstream of TSS. Some transcription factors have strongly conservation in every region we analyzed. These binding site distribution family-specific in transcription factors are helpful to predict which region contains the transcription regulation information so that a better algorithm for the discovery of transcription factor binding sites is possible.