Dinh/Dinh 2012/NOTES/2012-1-4
Jump to navigation
Jump to search
iPSC Epigenetics[edit]
Sergio's iPSCs[edit]
- Checked the localization of the CpGs found in shared aberrantly methylated genes:
File:Idiograph-sharedGenes-CGs.txt File:Idiograph-sharedGenes-CGs.png **Note: background are EST density by number.
- Shared aberrantly methylated genes tend to be away from the centrimere region.
iPSCs from Yang Xu's lab[edit]
- These iPSCs were reprogrammed using episomal vectors.
- First, I checked the 'reprogramming completeness' of different iPSC lines (including ones from whole genome bisulfite sequencing). To do this, I limited the sites to only CGs with no known disrupting SNP AND were not found to be differentially methylated between Hues1, Hues3, and Hues8 (methylation data from BSPP sequencing only).
- Below are the different pluripotent lines included for this analysis:
Source | Pluripotent lines | Cell type derived | Vectors | Method | #CpGs |
Xu lab (BSPP) | 218HFF211 | fetal liver fibroblasts | OCT4,SOX2,NANOG,LIN28A | Episomal | 286691 |
Xu lab (BSPP) | 218HFF411 | fetal liver fibroblasts | OCT4,SOX2,NANOG,LIN28A | Episomal | 266433 |
Xu lab (BSPP) | 2HFF8d_10 | fetal liver fibroblasts | OCT4,SOX2,NANOG,LIN28A | Episomal | 308102 |
WGBS | ADS-iPSC | adipose derived stem cells | OCT4, SOX2, KLF4, cMYC | Retroviral | 477356 |
Church lab (BSPP) | PGP1-iPSC | human adult fibroblasts | OCT4, SOX2, KLF4, cMYC | Retroviral | 395985 |
WGBS | FF-iPS-19-11 | foreskin fibroblasts | OCT4, SV40LT, NANOG, LIN28, KLF4 | Episomal | 291927 |
WGBS | FF-iPS-19-7 | foreskin fibroblasts | OCT4, SV40LT, NANOG, LIN28, KLF4 | Episomal | 316709 |
WGBS | FF-iPS-6-9 | foreskin fibroblasts | OCT4, SV40LT, NANOG, LIN28, cMYC, KLF4 | Episomal | 322250 |
WGBS | IMR90-iPS | fetal lung fibroblasts | OCT4, SOX2, NANOG, LIN28A | Lentiviral | 284770 |
WGBS | H9 | human embryonic stem cells | NA | NA | 266495 |
WGBS | H9 Laurent | human embryonic stem cells | NA | NA | 408474 |
WGBS | H1 | human embryonic stem cells | NA | NA | 305459 |
Xu lab (BSPP) | H1 | human embryonic stem cells | NA | NA | 400577 |
Xu lab (BSPP) | H3 | human embryonic stem cells | NA | NA | 305826 |
Xu lab (BSPP) | H8 | human embryonic stem cells | NA | NA | 454397 |
File:Xu dendrogram.png
- Next, I determined the sites which were aberrantly methylated using a Fisher Exact Test with Benjamini-Hochberg FDR of 0.01. Aberrant methylation were determined at sites where iPSCs were different from the combined Hues1, Hues3, and Hues8.
for f in mf_*.txt do ./getCpgChsq+DiffMatrix_DD_*.pl $f 10 0.2 0.01 H1H3H8_filteredsites_noCG-SNP > muir.$f.dmTable_noSNP awk ' function abs(value) { return (value<0?-value:value) }; { if($2==1 && $3==1 && abs($4-$5) >= 0.2 && ( ($4<0 && $5>=0) || ($4>0 && $5<=0) )) print $0,"\tMEMORY_INCOMPLETE"; //$2 is result from test vs hESC and $3 is result from test vs somatic source if($2==1 && $3==0 && abs($4-$5) >= 0.2 && abs($5) < 0.2 && abs($4) > abs($5) ) print $0, "\tMEMORY_COMPLETE"; if($2==1 && $3==1 && ( ($4<0 && $5<0) || ($4>0 && $5>0) )) print $0, "\tMUTATION"; if($2==0 && $3==1 && abs($4)<0.2) print $0,"\tPLURIPOTENCY"; } ' muir.$f.dmTable_noSNP > muir.$f.memmuTable_noSNP done;
- The aberrantly methylated sites were further limited to those which have <0.1 standard deviation between all of H1 (WGBS), H9 (WGBS), H9_Laurent (WGBS), Hues1, Hues3, and Hues8
- I uploaded the lists to [GREAT | http://www.great.stanford.edu] and obtained genes list for each of the 3 iPSC lines from Xu's lab.
- I took the intersection of the three lists to obtain shared methylation memory and shared methylation mutation gene lists.
File:Xu shared mutation.txt
- 1/5/11 - I forgot to separate out the genes associated with chrXYM
- BC018860 should be removed. (Shared mutated methylated on chrM in all three iPSCs.
File:Xu shared memory.txt
- MAGEE1 and ZDHHC15 should be removed. (Shared memory methylated in all three iPSCs.)
- Affy Expression Data, I got expression data from two samples, HuSCID and SCID.
- First, I uploaded the shared memory/mutation gene lists to DAVID and found all associated Affymetric probe IDs for each gene.
- Next, I annotated each probe as memory/mutation/normal.
- I divided the probes list into two: upregulated in HuSCID or downregulated in HuSCID. Additionally, I required that upregulated probes in HuSCID must NOT have detection value of 'A' (absence) and downregulated probes in HuSCID must NOT have detection value of 'A' (absence) in SCID.
- Next I calculated the fold change and rank them.
- Here are the list of aberrantly methylated genes and their fold-change in expression value in HuSCID/SCID (tumor samples). File:AberrantlyMethylatedGenesProbesExpression.xls