Changes between Version 59 and Version 60 of SnpCallingPipeline

Timestamp:

Jan 24, 2011 5:45:54 PM (15 years ago)

Author:

laurent

Comment:

--

SnpCallingPipeline

@@ -106 +106 @@
 == List of steps ==
 [[TOC(SnpCallingPipeline/ReferencePreparation,SnpCallingPipeline/AlignmentAndCleaning,SnpCallingPipeline/VariantCalling,inline,noheading)]]
+
+== Quality Control ==
+The current important values discussed for the quality control along with their thresholds are the following:
+* RawData
+** FastQC report (per mate of the pair)
+*** Manual look at files and check:
+**** Avg Quality per read > 30
+**** Num sequences ~60Mio
+**** Sequence quality should look OK
+* Alignment (per lane)
+** Picard Alignment Summary Metrics
+*** %Purified reads aligned > 90%
+*** Purified High Quality Error Rate < 1%
+*** Purified reads aligned > 150Mio
+** Picard GC Bias Metrics
+*** GC Curve should look OK
+*** Median GC% windows between 30 and 40
+*** Avg Mean Base Quality should be OK
+** Picard Insertsize Metrics
+*** Peak should be ~500
+*** Peak should be narrow
+*** Should have few outliers
+** Picard BAM Index Stats
+*** Should be uniform by Chromosome
+** GATK or Picard (currently testing) Coverage Metrics
+*** Should correspond to a Poisson curve with peak at 12x
+** Picard Mark Duplicates
+*** %duplicates between 5% and 8%
+* Recalibration
+** GATK Analyze Covariate
+*** No output currently; should revisit when working
+** Picard Quality by Cycle
+*** To be determined once data is produced
+** Picard Quality Distribution
+*** To be determined once data is produced
+* Initial SNP Calling
+** To be determined once data is produced and analyzed. A first basis for it should be derived from the difference between chipdata and sequence data and the %of SNPs found in dbSNP.
+