Athurva Gore/LabNotes/2009-8-20: Difference between revisions
Jump to navigation
Jump to search
>Ajgore (New page: {{ AGLabEntry|2009-8-19|2009-8-21 }} =False Positive SNPs= * Currently checking to see if clonal reads could be influencing false positive SNPs. * Added a step to variant caller that remo...) |
>Ajgore |
||
(One intermediate revision by the same user not shown) | |||
Line 5: | Line 5: | ||
* Added a step to variant caller that removes clonal reads (rmdup) | * Added a step to variant caller that removes clonal reads (rmdup) | ||
* Testing on HL003_merged "high false positive" set. | * Testing on HL003_merged "high false positive" set. | ||
** Removing clonal reads does not have any effect; downsampling seems to remove them in a similar way | |||
** No advantage gained. | |||
* MAQ, however, does something more interesting | |||
** MAQ only calls 2500 SNPs from HL003 NA12878 set. | |||
** Most of them are at dbSNP locations (only 300 not) | |||
** MAQ is being much more conservative somehow | |||
*** Taking into account mapping quality? What about taking into account neighboring consensus quality? | |||
=Probe Design= | |||
* Need to redesign two sets: | |||
** TDMR Set | |||
** CpG-SNP Set | |||
* Both should have H1+Target+H2 constrained to 105 | |||
* Also, adjust P2MIP script to have a maximum combined arm size of 45. | |||
* Allows sequencing without library construction |
Latest revision as of 23:18, 20 August 2009
[edit]
False Positive SNPs[edit]
- Currently checking to see if clonal reads could be influencing false positive SNPs.
- Added a step to variant caller that removes clonal reads (rmdup)
- Testing on HL003_merged "high false positive" set.
- Removing clonal reads does not have any effect; downsampling seems to remove them in a similar way
- No advantage gained.
- MAQ, however, does something more interesting
- MAQ only calls 2500 SNPs from HL003 NA12878 set.
- Most of them are at dbSNP locations (only 300 not)
- MAQ is being much more conservative somehow
- Taking into account mapping quality? What about taking into account neighboring consensus quality?
Probe Design[edit]
- Need to redesign two sets:
- TDMR Set
- CpG-SNP Set
- Both should have H1+Target+H2 constrained to 105
- Also, adjust P2MIP script to have a maximum combined arm size of 45.
- Allows sequencing without library construction