Kun:LabNotes/ExonomeSeq/2009-4-1

From ZhangLabWiki
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Data analysis of shotgun sequencing data[edit]

  • Two libraries were made by Alice. The capturing reactions were done with Jan09 #1-3 probe set (150k) on PGP1.
  • s3 (PGP1-CTRL): 9707995 raw reads, 40bp, 7655774 (79%) mappable (SOAP2 to hg18)
  • s4 (PGP1-Shaked): 9424752 raw reads, 40bp, 7561445 (80%) mappable (SOAP2 to hg18)

Questions to answer[edit]

Was the library construction successful?[edit]

It seems pretty good. Here is the distribution of reads that mapped to the same location. File:2009-03-30-clonality check.png

What is the level of representation bias?[edit]

It is slightly higher than the bias in end-sequencing analysis, but not too much. The library construction could be improved. Library-free sequencing would be better.

 File:2009-03-30-AbsoluteCoverageDist.png File:2009-03-30-RelativeCoverageDist.png

What is the fraction of on-target reads?[edit]

  • Read mapping using Soap2 with the default setting.
    • 19,132,747 total reads.
    • 15,217,219 mapped.
  • I used this Perl script to generate the following statistics:
                       Genome covered(bps)   Sequenced(bps)      
   Mapped reads:	44,333,324	      608,688,760
   On-target reads:	15,064,164	      499,303,049 (33x)  82%
   Near-target	reads:	 4,005,655	      93,306,387  (23x)  15%

What is the fraction of reads that came from linker sequences or capturing arms?[edit]

How many SNPs were identified? What is the false positive rate?[edit]

  • I've tried both Maq and Soap on either the full genome template or the captured targets only. SNP calling was based on the minimal sequencing coverage of 8x and the minimal quality score of 30.
                 Full genome       Targets only
      Maq         10,655              10,202
      SOAP                            10,698
   File:Comparison of SNP calling strategies.png