基本HTML版本
Computational Molecular Biology 2014, Vol. 4, No. 10, 1-17
http://cmb.biopublisher.ca
6
Table 2 Number of partners for all H3K4me localization types
Degree
Gene number with PPI annotation/All gene
me1me2
10.98±20.65
264/688
me1me2me3
10.71±15.94
123/332
me1me3
11.99±21.97
182/451
me2me3
11.01±18.69
401/1021
Average*
11.07±19.57
774/1986
Single
9.38±16.13
4008/10222
Note: * Significant (
p
=0.02)
2.3 Functional assessment of co-localized peaks
using GO terms
Genes associated with co-localized peaks in TPRs
(upstream 1k and downstream 2k around TSSs) may
be enriched with specific gene functions compared
with controls, such as transcription factor binding
regulation. To verify this, GO enrichment analysis for
genes associated with co-localized peaks in TPRs was
performed. We used the DAVID system, which can
identify over-represented GO terms for a set of genes
(full results in Table 3). A Venn diagram for
representing overlapping GO terms was shown in
Figure 3A. As was shown in Figure 3A, 11 GO terms
are shared by four co-localization types. These terms
were supposed to be associated with general cellular
component. Among them, two were related with
protein binding, three were related with intracellular
terms, and another three were related with organelle.
Among the four co-localization types, me1me3 was a
prominent type in that it has been previously reported
that mono-methylation together with marginal
tri-methylation is the mark of enhancer
. Furthermore, it was also supported by a
recent study, in which Transcription Factor Binding
Sites (TFBSs) were found to be associated with
H3K4me1 and H3K4me3 co-localization
. In this study, for all six associating GO
terms listed in Table 3, all were directly or indirectly
related with transcriptional regulation. Interestingly,
we observed that the lengths of genes associating with
me1me3 (length = 46,521) and me1me2 (length =
45,961) were significantly shorter than the control
group (length = 51,615). Our results highlighted the
potentially new regulatory role of me1me3 for short
genes. In contrast to me1me3, the average gene length
was the largest for me2me3 (length = 59,568).
Consistent with a previous study in
Arabidopsis
thaliana
, our result also
highlighted me3 mark was a faithful guard for long
genes and me1 mark was a guard for short genes.
Table 3 Exclusively enriched GO terms in co-localization types of me1me3, me1me2me3 and me2me3
Localization GO term
Localization GO term
me1me3
Transcription, DNA-dependent
me2me3
Cellular component assembly
me1me3
Regulation of metabolic process
me2me3
Macromolecular complex assembly
me1me3
Transcription regulator activity
me2me3
Cell development
me1me3
Regulation of cellular metabolic process
me2me3
Transferase activity, transferring
phosphorus-containing groups
me1me3
RNA biosynthetic process
me2me3
Endosome membrane
me1me3
Positive regulation of biological process
me2me3
Endosomal part
me1me2me3 Cell
me2me3
Kinase activity
me1me2me3 Protein kinase cascade
me2me3
Transferase activity
me1me2me3 Transcription repressor activity
me2me3
Cell cycle
me2me3
Protein import into nucleus
me2me3
Protein targeting
me2me3
Endoplasmic reticulum membrane
me2me3
Enzyme binding
me2me3
Positive regulation of programmed cell death
me2me3
Endoplasmic reticulum part
me2me3
Nuclear import
me2me3
Positive regulation of apoptosis
me2me3
Intracellular protein transport
me2me3
Regulation of a molecular function
me2me3
Nuclear envelope-endoplasmic reticulum network me2me3
Nucleosome
me2me3
Post-translational protein modification
me2me3
Protein import
me2me3
Cell cycle process