Dinh 2011/NOTES/2011-11-30
Jump to navigation
Jump to search
UCLA Sib-pair Analysis[edit]
- Sib-pair data: for every affect indiv take methyl frequency and substract the methyl frequency of an unaffected sibling
- Null-pair data: for every inter-family pair, take the methyl frequency of mother and substract the methyl frequency of the father or take the methyl frequency of child and substract methyl frequency of parent
- User perl's Statistics::TTest module. Test whether the means are significantly different by p<=0.05 between the sib-pair set and the null pair set
./cpgMatrixSibPairTTest.pl sample_info.txt sites_KW_unsignif0.10n5 > fam_filtered_sibpairTstat.LOG &
- File:CpgMatrixSibPairTTest.txt
- Outputs: File too big, see /home/ddiep/UCLA_Analysis/sibpairTstat.KWunsignif0.10n5.matrix (meangenemachine)
- Hierarchical clustering at sib pair significant sites: File:SibpairHclust UCLA.pdf
There are two branches and all affected individuals are on one branch (some unaffected individuals are also on same branch).
UCLA overlap MPO and Sib-pair[edit]
- MPO analysis by Noi
- Overlap mpo-noSNP_heritable-rd1 with sibpair results:
perl matrices/getSites.pl matrices/mpo-noSNP_heritable-rd1 < sibpairTstat.KWunsignif0.10n5.out > mpo_sibpair_kwunsignif0.10n5.out
- Got 991 sites
- Make BED files:
awk '{print $1}' mpo_sibpair_kwunsignif0.10n5.out | sed 's/:/\t/g' | awk '{print $1"\t"$2-1"\t"$2;}' > mpo_sibpair_kwunsignif0.10n5.BED File:Mpo sibpair kwunsignif0.10n5.BED.txt awk '{print $1}' matrices/methyl_min76_minSTD0.1_rd1 | sed 's/:/\t/g' | awk '{print $1"\t"$2-1"\t"$2;}' > cpg.bg.BED File:Ucla min76 std0.1 rd1 cpg.bg.BED.txt
- GREAT analysis
File:Mpo-sibpair-UCLA-genes.txt File:Mpo-sibpair-UCLA-ontologies.txt
UCLA Genetic Distance[edit]
- Make a list file: (meangenemachine: /home/ddiep/UCLA_Analysis/)
ls hg19SNPfiltered/*.filtered > list_snp awk '{print "123"$0"123\t"$0;}' list_snp | sed 's/123hg19SNPfiltered\///g' | sed 's/.fwd.pileup.snp.filtered123//g' > snp_list
- Make tped and tfam files:
./getTPED.pl hg19SNP.tfam hg19SNP.tped < snp_list
- Calculate genetic distance:
./tped2GeneticDistance.pl hg19SNP.tfam hg19SNP.tped > hg19.dist
There appears to be some sample mislabeling. GK0050-005a appears to be GK0277-002 ( male to female labeling ) GK0277-002 appears to be GK0344-004 ( female to male labeling ) GK0344-004 appears to be GK0050-005a ( male to male labeling )
- We can correct 2/5 samples which appear to have the wrong sex labeling.
- Leftover questionable samples are: GK0210-002, GK0210-001, and GK0396-004.
- How should we deal with those samples ? Switch them back ?
GK0210-001 (labeled as father) could be switched with GK0210-002 (labeled as mother) GK0396-004 (labeled as male child) appears to be female child.