Dinh 2011/NOTES/2011-12-27

From ZhangLabWiki
Jump to navigation Jump to search

Probes Resubsetting[edit]

Sub1-7 (DMR220K)[edit]

  • 20,438 probes did not capture with normalization and in multiple capture experiments.
 File:Sub1 7 0capture.txt
 Out of the 20,438: 1,039 captured in the rebalancing experiment so I will put them back (these were removed from the 0 capture file above).
 The reason for placing them back is because these probes were able to capture and maybe we lost them during probes preparation. Hopefully, we can get these back after re-synthesis.
  • 99,054+1,039 probes to resubset
 File:Sub1 7 Capture.txt - I wasn't able to upload this file
  • For Sub1_7 (DMR220K) I have only 1 capture experiment to resubset with.
 File:Sub1 7 efficiencyPlot.png
 Subsetting scores (normalized counts, not log scores):
        A > 200 (Subset 1) File:Sub1 7.s1.txt 3544 (#probes)
  200=> B >  20 (Subset 2) File:Sub1 7.s2.txt 30213
   20=> C > 5.4 (Subset 3) File:Sub1 7.s3.txt 29818
  5.4=> D > 0 (Subset 4)   File:Sub1 7.s4.txt 24747
      E = 0 (Unk)          File:Sub1 7.s5.txt 11771
 File:Sub1 7 resubset1.png
 File:Sub1 7 resubset2.png

Exp1-3 (Expansion probes)[edit]

  • 32,144 probes did not capture with normalization and in multiple capture experiments.
 File:Exp1 3 0capture.txt
  • 106,071 probes to resubset
 File:Exp1 3 Capture.txt - I wasn't able to upload this file
  • For Exp1_3 (Expansion set) I have multiple capture experiment to work with (also in 10 different tissues). I used the average normalized read counts to subset.
 File:Exp1 3 efficiencyPlot.png
 Subsetting scores (normalized counts, not log scores):
         A > 35 (Subset 1)    File:Exp1 3.s1.txt  3428 (#probes)
   35 => B > 5 (Subset 2)     File:Exp1 3.s2.txt 23849
    5 => C > 1.3 (Subset 3)   File:Exp1 3.s3.txt 28290
  1.3 => D > 0.25 (Subset 4)  File:Exp1 3.s4.txt 22440
  0.25 >= E (performs poorly) File:Exp1 3.s5.txt 28064
 File:Exp1 3 resubset1.png
 File:Exp1 3 resubset2.png

Notes[edit]

  • All reads counts were normalized to per 1 million mapped reads.
  • From the rebalancing experiment, I was able to find 8,796 probes involved in off-target capture or inaccurate mapping of reads to probes.
  • The off-target capture read counts were between 1-327, with an average of ~ 1.6 reads per 1 million mapped reads.
  • It would be helpful to learn more about the reads which were not able to map to any probes, especially for Sub5,Sub6, and Sub7. We know that DMR330K after the first normalization have >90% specificity.
  • How did not using Sub6 and Sub7 affected specificity?
  • Closer examination of sensitivity and specificity calculation from a few months ago,
The recent PE100 library was from the UPENN set, and was captured with Sub6 and Sub7 => slightly lower specificity
The recent PE110 library as well as those in HL094 was not captured with Sub6 and Sub7 => slightly higher specificity, with the exception of PGP1F (may have low DNA quality).
  • Thus, DMR330K has high specificity because the off-target capturing probes were also the least efficient probes.
  • LC4K probes probably have low specificity but comparable sensitivity because many of the probes are efficient and because of deep sequencing coverage.
' Total target size Total mappable bases Total on target bases Total target covered >=1X Sensitivity (%target covered) Specificity (%mappped base on target)
LC4K
H1-54C 251797 168941334 89608268 185056 0.73494124234999 0.53041056252107
H1-52C 251797 189704106 78915250 187795 0.74581905264956 0.41599125956715
PGP1iPS-54C 251797 210319026 131291469 185174 0.73540987382693 0.62424912998599
PGP1iPS-52C 251797 257710476 124280612 188753 0.74962370480983 0.48224897151639
PGP1F-54C 251797 301811294 159092545 188938 0.7503584236508 0.52712588349991
PGP1F-52C 251797 301526024 121979569 193326 0.76778516026799 0.40454076693559
DMR330K 0.49742776233768
Recent library-free PE110 34394092 2340766083 2254282508 24118127 0.70122877498845 0.96305330309248
*Recent library-free PE100 34394092 1687630717 1582327158 25050732 0.72834404234309 0.93760272437611
HL094-H1 34394092 2340766083 2254282508 24118127 0.70122877498845 0.96305330309248
HL094-PGP1F 34394092 1019718323 934477017 13888586 0.4038073166752 0.9164070076242
HL094-PGP1iPS 34394092 1900359617 1830971383 25334157 0.73658455644068 0.96348678777465