Kun:LabNotes/Haplotyping/2011-9-7

From ZhangLabWiki
Revision as of 17:59, 10 September 2011 by KunZhang (talk | contribs) (→‎Variant calling and visualization of haplotypes)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Haplotyping data analysis[edit]

  • Rui&Matt lysed GM12878 cells, and made dilution to ~8 chrosomes per aliquot, then performed tagmentation followed by PCR with indexing primers. Rui prepared 24 libraries (one failed) and Matt prepared 12 libraries (one failed). Each mixture of 12 libraries was sequenced in two lanes (HL101, 60bp+index): Lab notes.

Mapping[edit]

  • I processed 12 data sets in a batch.
  less /home/kunzhang/SeqStore/110902_HL101/s_1_Indx1.txt.gz /home/kunzhang/SeqStore/110902_HL101/s_2_Indx1.txt.gz > s12_Indx1.fq
  less /home/kunzhang/SeqStore/110902_HL101/s_1_Indx2.txt.gz /home/kunzhang/SeqStore/110902_HL101/s_2_Indx2.txt.gz > s12_Indx2.fq
  less /home/kunzhang/SeqStore/110902_HL101/s_1_Indx3.txt.gz /home/kunzhang/SeqStore/110902_HL101/s_2_Indx3.txt.gz > s12_Indx3.fq
  less /home/kunzhang/SeqStore/110902_HL101/s_1_Indx4.txt.gz /home/kunzhang/SeqStore/110902_HL101/s_2_Indx4.txt.gz > s12_Indx4.fq
  less /home/kunzhang/SeqStore/110902_HL101/s_1_Indx5.txt.gz /home/kunzhang/SeqStore/110902_HL101/s_2_Indx5.txt.gz > s12_Indx5.fq
  less /home/kunzhang/SeqStore/110902_HL101/s_1_Indx6.txt.gz /home/kunzhang/SeqStore/110902_HL101/s_2_Indx6.txt.gz > s12_Indx6.fq
  less /home/kunzhang/SeqStore/110902_HL101/s_1_Indx7.txt.gz /home/kunzhang/SeqStore/110902_HL101/s_2_Indx7.txt.gz > s12_Indx7.fq
  less /home/kunzhang/SeqStore/110902_HL101/s_1_Indx8.txt.gz /home/kunzhang/SeqStore/110902_HL101/s_2_Indx8.txt.gz > s12_Indx8.fq
  less /home/kunzhang/SeqStore/110902_HL101/s_1_Indx9.txt.gz /home/kunzhang/SeqStore/110902_HL101/s_2_Indx9.txt.gz > s12_Indx9.fq
  less /home/kunzhang/SeqStore/110902_HL101/s_1_Indx10.txt.gz /home/kunzhang/SeqStore/110902_HL101/s_2_Indx10.txt.gz > s12_Indx10.fq
  less /home/kunzhang/SeqStore/110902_HL101/s_1_Indx11.txt.gz /home/kunzhang/SeqStore/110902_HL101/s_2_Indx11.txt.gz > s12_Indx11.fq
  less /home/kunzhang/SeqStore/110902_HL101/s_1_Indx12.txt.gz /home/kunzhang/SeqStore/110902_HL101/s_2_Indx12.txt.gz > s12_Indx12.fq
  fastq2variants.pl s12_Indx1.fq 60 > s_12_Indx1.log
  fastq2variants.pl s12_Indx2.fq 60 > s_12_Indx2.log
  fastq2variants.pl s12_Indx3.fq 60 > s_12_Indx3.log
  fastq2variants.pl s12_Indx4.fq 60 > s_12_Indx4.log
  fastq2variants.pl s12_Indx5.fq 60 > s_12_Indx5.log
  fastq2variants.pl s12_Indx6.fq 60 > s_12_Indx6.log
  fastq2variants.pl s12_Indx7.fq 60 > s_12_Indx7.log
  fastq2variants.pl s12_Indx8.fq 60 > s_12_Indx8.log
  fastq2variants.pl s12_Indx9.fq 60 > s_12_Indx9.log
  fastq2variants.pl s12_Indx10.fq 60 > s_12_Indx10.log
  fastq2variants.pl s12_Indx11.fq 60 > s_12_Indx11.log
  fastq2variants.pl s12_Indx12.fq 60 > s_12_Indx12.log
  • Then I used samtools to get the read counts for each data set:
 /home/kunzhang/softwares/samtools-0.1.12a/samtools view -c  s12_Indx1.fq.bowtie.sorted.bam
 /home/kunzhang/softwares/samtools-0.1.12a/samtools view -c  s12_Indx1.fq.bowtie.sorted.unique.bam
 File:Haplotyping HL101 library complexity.png
    • We got an average of 3.5 millions reads (median 3.5M) per library, and the average unique reads per library is ~250k (median 117k). This is roughly twice as high as the 12 libraries Rui did in June, which could be due to the use of protease digestion after tagmentation.
    • With the combination of 34 libraries (equivalent to 6 cells), we have 8.5 millions of unique 60bp reads, which cover 500Mb (or 1/6) of the human genome.
    • If we obtain PE100bp reads, we would cover roughly half of the human genome with the same set of libraries.
    • To obtain 1X coverage of the genome, we need PE100bp reads on ~80 of such libraries.

Variant calling and visualization of haplotypes[edit]

 File:RuiA Indx1 idioGraph.pngFile:RuiA Indx2 idioGraph.png
 File:RuiA Indx3 idioGraph.pngFile:RuiA Indx4 idioGraph.png
 File:RuiA Indx5 idioGraph.pngFile:RuiA Indx6 idioGraph.png
 File:RuiA Indx7 idioGraph.pngFile:RuiA Indx8 idioGraph.png
 File:RuiA Indx9 idioGraph.pngFile:RuiA Indx10 idioGraph.png
 File:RuiA Indx11 idioGraph.pngFile:RuiA Indx12 idioGraph.png
 File:RuiB Indx1 idioGraph.pngFile:RuiB Indx2 idioGraph.png
 File:RuiB Indx3 idioGraph.pngFile:RuiB Indx4 idioGraph.png
 File:RuiB Indx5 idioGraph.pngFile:RuiB Indx6 idioGraph.png
 File:RuiB Indx7 idioGraph.pngFile:RuiB Indx8 idioGraph.png
 File:RuiB Indx10 idioGraph.png
 File:RuiB Indx11 idioGraph.pngFile:RuiB Indx12 idioGraph.png
 File:Matt Indx1 idioGraph.pngFile:Matt Indx2 idioGraph.png
 File:Matt Indx3 idioGraph.pngFile:Matt Indx4 idioGraph.png
 File:Matt Indx5 idioGraph.pngFile:Matt Indx6 idioGraph.png
 File:Matt Indx8 idioGraph.png
 File:Matt Indx9 idioGraph.pngFile:Matt Indx10 idioGraph.png
 File:Matt Indx11 idioGraph.pngFile:Matt Indx12 idioGraph.png
  • Some really promising ones include RuiA: Indx5, 6, 8; RuiB: Indx4, 8; Matt: Indx3, 4, 5, 7.
  • The question is how to call haplotypes on samples without references.

How to further improve?[edit]

  1. Pre-amplification using a single primer targeting the transposon recognition sequence, so that all A-A, A-B, B-B fragments can be captured.
  2. Use metaphase-blocked cells to increase the average fragment size of DNA molecules.
  3. Develop a QPCR assay to estimate the complexity of the libraries without Illumina sequencing. The key is to establish a reference library of known quantity, dilute the library to different levels (250k molecules, 500k molecules, 1M molecules, 2M molecules), and include these diluted libraries as the standards for the post-tagmentation PCR.