Learn more about Picard's Mark Duplicates
If you want to run Picard's Mark Duplicates on Lonestar, do roughly this:
module load picard # "picard" is not the name of the tool; picard is a collection of java .jar files that do things, so we have to figure out where they are and which ones they are, like this: ls $TACC_PICARD_DIR # Now to run Mark Duplicates, do something like: java -Xmx2g -jar $TACC_PICARD_DIR/MarkDuplicates.jar INPUT=all.bowtie.sorted.bam OUTPUT=all.bowtie.sorted.dedup.bam METRICS_FILE=all.bowtie.sorted.dedup.bam.metrics ASSUME_SORTED=TRUE
The output metrics file is required and worth screening before you go further.
Learn more about picard on Fourierseq
Picard is also installed on fourierseq at /usr/local/genome/bin and the jar files are in fourierseq user's paths.
General syntax for using picard jar files:
java -Xmx4g -jar /usr/local/genome/bin/jarfilename
Find picard document here