Anexo i Entrevistas a estudiantes de preescolar a quinto de primaria
Fuente 30 Sesión de grupo con estudiantes Servicio de las clases de ERE
To prioritise the most disease-relevant variants, four levels of filtering were applied to significantly associated loci. Figure 4.9 highlights the four filtering stages, together with the number of CpGs progressing through each level. In the first instance,
meQTLs with a –log10(p-value) below the adjusted significance threshold of 10 were
removed from further analysis, leaving the 111 most significant associations, with forty-eight unique CpGs from the risk loci approach and sixty-six from the variable methylation approach, with three CpGs overlapping between the variable
methylation and risk loci approaches.
To maintain the focus on CpG-SNPs, loci lacking a CpG-SNP were removed. Ninety- nine CpGs remained, sixty-six from the variable methylation approach and thirty- three from the risk loci approach. For each of these ninety-nine CpGs the function of the nearest gene was examined. While CpG methylation and other regulatory motifs do not always affect expression at the most proximal gene, instead influencing expression at distal regions (in trans), current analysis tools make it extremely
challenging to accurately examine regulatory function outside proximal genes. Genes with a reported role as a tumour suppressor or oncogene, a role in proliferation, cell cycle regulation, angiogenesis or gene regulation were included. Illumina’s genomic annotation on the methylation array was used to determine the most proximal gene, namely if CpGs were located within a gene or within 1,500bp of a transcription start site. These annotations were then confirmed using the UCSC genome browser. CpG sites that were not annotated to a gene on the methylation array were excluded from further analysis. Thirty-seven CpG sites remained after this filtering, and are
presented in Table 4.3 together with a subset of the annotation information and reported functionality of the nearest gene.
Table 4.3 The 37 most significant associations from all approaches.
CpGs are filtered by significance of p-value, presence of CpG-SNP, absence of probe SNP and gene function. CpG sites prioritised for validation are highlighted in orange with three CpG sites prioritised from both approaches highlighted in yellow.
A) meQTLs identified throughstandard deviation CpG Name In other variable
approach CpG-SNP Gene Gene Function
1 cg13387643 no rs284310 CASZ1 Zinc finger transcription factor, may function as tumor suppressor
2 cg09084244 no rs1109559 CDK2AP1 Possible regulatory role in DNA replication. Forms core subunit of NURD complex; epigenetically regulates embryonic stem cell differentiation. Associated with oral cancer
3 cg25013753 no rs1051508 ARHGAP22 Regulates cell motility & angiogenesis. May be involved in transcription regulation via interaction with VEZF1 4 cg08210706 no rs10135403 SERPINA5 Inhibits urinary-type plasminogen activator-dependent tumor cell invasion and metastasis
5 cg00231519 no rs36101953 C10orf46 Cell cycle associated protein capable of promoting cell proliferation
6 cg02978201 no rs737008 PRM1 Protamines substitute for histones in the chromatin of sperm during the haploid phase of spermatogenesis
7 cg00345083 no rs7517857 AJAP1 Plays a role in cell adhesion and cell migration
B) meQTLs identified through 95% Reference Range CpG Name variable In other
approach CpG-SNP Gene Gene Function
9 cg20592836 yes rs2378256 TP53INP2 Dualrole as transcription factor & in autophagy
10 cg02464073 no rs1721 ITGB2 leukocyte adhesion deficiency & gastrointestinal carcinoma Important role in immune response. Defects in gene cause
11 cg08146865 yes rs3197223 NME6 Inhibitor of p53-induced apoptosis
12 cg23698271 no rs11199030 TIAL1 RNA-binding protein, regulates various activities including translational control, splicing & apoptosis
13 cg21927991 yes rs5025124 ZFAT Puatively binds DNA & functions as a transcriptional regulator
involved in apoptosis and cell survival
14 cg07240846 yes rs10906142 CAMK1 nuclei, possible involvement in apoptosis of erythroleukemia Activates transcription factor CREB1 in hippocampal neuron cells
15 cg06330797 yes rs7357046 RPS6KA2 Implicated in cell growth & differentiation, may function as tumor suppressor in ovarian cancer 16 cg19360212 yes rs11200296 NSMCE4A recombination, required for telomere maintenance, involved in Involved in DNA double-strand breaks by homologous
positive regulation of response to DNA damage stimulus
17 cg05331763 no rs79974293 FOXK2
Related pathways: Cell Cycle / Checkpoint Control and Wnt / Hedgehog / Notch. GO annotations include sequence-specific
DNA binding transcription factor activity & magnesium ion binding
18 cg01891583 yes rs2304466 USP7 May induce p53/TP53-dependent cell growth repression & apoptosis
acute myelomonocytic leukemia
20 cg11251367 yes rs12403072 FMN2 Role in organization of cytoskeleton & cell polarity. Involved in responses to DNA damage
21 cg09993319 yes rs7898151 MGMT Involved in Cell Cycle & DNA-methyltransferase activity
22 cg04610028 yes rs2967607 RAB11B Involved in regulating exocytotic and endocytotic pathways
23 cg05338731 no rs5751591 RAB36 Associated with rhabdoid cancer
24 cg02658043 yes rs7465214 NRBP2 Associated with medulloblastoma
25 cg26365090 yes rs11700304 TOX2 RNA polymerase II transcription factor binding. Putative transcriptional activator 26 cg17662493 yes rs6006744 SMC1B Involved in chromatid cohesion and DNA recombination during meiosis and mitosis
27 cg03796003 yes rs117229426 KCTD5 Associated with rectal neoplasm
C) meQTLs identified through the Tasmanian Familial Prostate Cancer Study
CpG Name Variable
Approach CpG-SNP Gene Gene Function
29 cg03036702 No rs4714482 FOXP4 Transcription factor, may play a role in tumors of the kidney & larynx
30 cg16995742 No rs7577630 COPS8 Important regulator in multiple signaling pathways
D) meQTLs identified through published familial prostate cancer studies
CpG Name Approach Variable CpG-SNP Gene Gene Function
cg11251367 Yes rs12403072 FMN2 Role in organization of actin cytoskeleton & cell polarity. Involved in responses to DNA damage, cellular stress & hypoxia
31 cg00069771 No rs17452776 C1orf57 Involved in purine metabolism
32 cg23209941 No rs12137417 DISC1-TSNAX DISC1: involved in neurite outgrowth & cortical development
33 cg07134368 No rs12034296 TSNAX-DISC1 sequences at breakpoint junctions of chromosomal translocations TSNAX: Interacts with DNA-binding protein that binds consensus
E) meQTLs identified through published prostate cancer GWAS
CpG Name Approach Variable CpG-SNP Gene Gene Function
34 cg23069046 No rs6920276 REXO2 Possible role in DNA repair, replication, recombination
cg03036702 No rs4714482 FOXP4 Transcription factor, implicated in kidney & larynx tumors
35 cg09349613 No rs72828989 CTBP2 One alternative transcript is a transcriptional repressor
To further prioritise prostate cancer relevant meQTLs, the presence of putative regulatory elements were examined 500 bp either side of the identified meQTLs. Again, the UCSC genome browser was used for this approach, specifically to determine the presence of regulatory elements such as transcription factor and miRNA binding sites, active histone and chromatin marks and DNase hypersensitivity regions. Concurrently considered, was the CpG density of each region, with regions containing too few CpGs (less than ten) excluded and regions containing dense CpG islands also excluded, as recent evidence suggests the greatest inter-individual variability occurs outside CpG islands, at shores and shelves. Utilising this filtering strategy, twenty-three CpGs were prioritised for further analysis and these are highlighted in orange in Table 4.3.
Figure 4.9 MeQTL filtering steps.
The five levels of filtering applied to the significant associations are presented