Genome-wide analysis of regions similar to promoters of histone genes

Abstract

Background: The purpose of this study is to: i) develop a computational model of promoters of human histone-encoding genes (shortly histone genes), an important class of genes that participate in various critical cellular processes, ii) use the model so developed to identify regions across the human genome that have similar structure as promoters of histone genes; such regions could represent potential genomic regulatory regions, e.g. promoters, of genes that may be coregulated with histone genes, and iii/ identify in this way genes that have high likelihood of being coregulated with the histone genes.

Results: We successfully developed a histone promoter model using a comprehensive collection of histone genes. Based on leave-one-out cross-validation test, the model produced good prediction accuracy (94.1% sensitivity, 92.6% specificity, and 92.8% positive predictive value). We used this model to predict across the genome a number of genes that shared similar promoter structures with the histone gene promoters. We thus hypothesize that these predicted genes could be coregulated with histone genes. This hypothesis matches well with the available gene expression, gene ontology, and pathways data. Jointly with promoters of the above-mentioned genes, we found a large number of intergenic regions with similar structure as histone promoters.

Conclusions: This study represents one of the most comprehensive computational analyses conducted thus far on a genome-wide scale of promoters of human histone genes. Our analysis suggests a number of other human genes that share a high similarity of promoter structure with the histone genes and thus are highly likely to be coregulated, and consequently coexpressed, with the histone genes. We also found that there are a large number of intergenic regions across the genome with their structures similar to promoters of histone genes. These regions may be promoters of yet unidentified genes, or may represent remote control regions that participate in regulation of histone and histone-coregulated gene transcription initiation. While these hypotheses still remain to be verified, we believe that these form a useful resource for researchers to further explore regulation of human histone genes and human genome. It is worthwhile to note that the regulatory regions of the human genome remain largely un-annotated even today and this study is an attempt to supplement our understanding of histone regulatory regions.

Figure 1

Bayesian network model of histone promoter structure.

Figure 2

Extraction of segments by scanning the genome using the CAAT-box PWM.

Figure 3

DPM’s 200,180 predictions on different chromosomes and their mappings with known gene transcripts.

Figure 4

Median chromosomal distance between intergenic predictions and their nearest gene.

Figure 5

Positional bias of predictions on mapped genes.

Figure 6

Distribution of predictions on mapped gene regions.

Figure 7

Distribution of prediction mapping with genes. GeneIDs (represented in the graph by their notional indices; refer Additional file 14 to see actual geneIDs) on the left have higher number of mapped predictions.

Figure 8

Distribution of CC between Predicted and Non-Predicted probe sets.

Figure 9

(a) Enriched GO term branch of nucleotide metabolism; (b) Enriched GO term branch of cell cycle regulation.

Figure 10

Top ranked network enriched by the predicted gene list with IPA.

Figure 11

Enriched functions with IPA.