example for using Picard removing duplicate reads?

[fabrice]

Does anybody can give me such example?

java -jar ~/bin/picard/MarkDuplicates.jar REMOVE_DUPLICATES=true METRICS_FILE=dup.txt INPUT=accepted_hits.bam OUTPUT=remove_accepted_hits.bam

Thank you very much.

[C.R.]

Hi,
I usually increase the RAM used, change the validation stringency and define a tmp dir.

My example will look like this:
java -Xmx4g -jar ~/bin/picard/MarkDuplicates.jar INPUT=accepted_hits.bam OUTPUT=remove_accepted_hits.bam METRICS_FILE=dup.txt VALIDATION_STRINGENCY=LENIENT REMOVE_DUPLICATES=true TMP_DIR=/tmp

[fabrice]

Thank you for your suggestions.
I am doing RNA-seq data analysis. I am still not very sure it is neccessary to remove the duplication or not.

[fabrice]

When I used picard to remove dulications, I used fastqc to check. It still have a high duplication level (63%). Why?

[chongm]

Maybe your library has low complexity?

[rosa_dentellare]

I'm having trouble with this MarkDuplicates. I hope someone could help. When I didn't include the "VALIDATION_STRINGENCY=LENIENT" I got an error saying "Mate Alignment start should be !=0 because reference name != *". But when I included it there's these error line that keeps changing extremely but its running. Is this normal?

[dariober]

Thank you for your suggestions.
I am doing RNA-seq data analysis. I am still not very sure it is neccessary to remove the duplication or not.

I think in RNAseq is normal to see reads piling up at the same position (i.e. duplicates), especially in highly expressed genes and/or with high sequencing depth. So I don't think it is recommendable to remove duplicates. But see also this thread for longer discussion http://seqanswers.com/forums/showthread.php?t=6854.

[vd4mindia]

I need some suggestion about the mark duplicates command using picard tool, can anyone tell me if in the input and out label we can specify path as well or not? as I want to run 6 samples together so I will use it as a script and execute all of them one by one so can I specify path in the input and out like below.

java -Xmx14g -jar /data/PGP/gmelloni/picard-tools-1.84/picard-tools-1.84/MarkDuplicates.jar INPUT=/test/exome/input/SRR062634.sorted.bam OUTPUT=/test/exome/results/SRR062634_marked.sorted.bam REMOVE_DUPLICATES=False METRICS_FILE=metricN.log ASSUME_SORTED=True VALIDATION_STRINGENCY=LENIENT

[dariober]

can anyone tell me if in the input and out label we can specify path as well or not?

java -Xmx14g -jar /data/PGP/gmelloni/picard-tools-1.84/picard-tools-1.84/MarkDuplicates.jar INPUT=/test/exome/input/SRR062634.sorted.bam OUTPUT=/test/exome/results/SRR062634_marked.sorted.bam REMOVE_DUPLICATES=False METRICS_FILE=metricN.log ASSUME_SORTED=True VALIDATION_STRINGENCY=LENIENT

I'm pretty sure you can specify the path, just try and see...

[vd4mindia]

Yes I have queued a script and its still running. Since am working from home my intenet is a bit weak so have to run script in cluster and cant asses the command line script so I was asking but it seems to work for me. Will keep posted if its successful or failed.