Downsampling a BAM file

[KathrineBL]

Hi ChIP-seq experts

I'm a newbie in the field of ChIP-seq data mining and need some help! I've sequenced several samples and mapped them using bowtie and everything looks fine so fare

But they differ somewhat in sequence depth which makes them difficult to compare - until now I've used coverageBed (bedtools) to find the read coverage around TSS and in my peak regions and then normalized the read count in these regions to sequence depth.

But for some of my future analysis it would be really nice if the BAM file was normalized to sequence depth - simply, I want to remove some random reads from one sample so it has the same amount of reads as my second sample... I've found that picard "DownsampleSam" should be able to do this, however I cannot get the programme to work on my (mac) computer.

I hope someone can help!!

BR, Kathrine

[ETHANol]

http://seqanswers.com/forums/showthread.php?t=16845

If you have no headers or you convert to BED the attached perl script should work as well. Warning: I'm not really sure what the script does, but it seems to work.

[pbluescript]

You could use bamtools random for this as well.
What errors is Picard giving you?

[KathrineBL]

Thank you so much for your input!!

I've tried the code, but i doesn't seem to work - I'm not that much up for converting it to a bed file as I need the BAM format later on.

I'e tried bamtools random, but keep on getting the same error message

bamtools random ERROR: could not load index data for all input BAM file(s)... Aborting.

My code line is as follows:

bamtools random -in Input_file.bam -out output_reduced.bam -n 1000000

The input bam file originates from the SAM file produced when mapping with bowtie - it is converted to BAM with "samtools view", and sorted with "samtools sort" - then I extract all mapped reads with "samtools view -b -F 4"

As you might can imagine I'm a newbie in this field and all help is very much appreciated!

[KathrineBL]

Regarding the Picard errors - I think it relates to the (mac) version of my java (which otherwise is up to date):

Exception in thread "main" java.lang.NoClassDefFoundError: jvm-argsCaused by: java.lang.ClassNotFoundException: jvm-args
at java.net.URLClassLoader$1.run(URLClassLoader.java:202)
at java.security.AccessController.doPrivileged(Native Method)
at java.net.URLClassLoader.findClass(URLClassLoader.java:190)
at java.lang.ClassLoader.loadClass(ClassLoader.java:306)
at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:301)
at java.lang.ClassLoader.loadClass(ClassLoader.java:247)

[pbluescript]

You need to index the bam file first. You can do this with bamtools:

bamtools index -in Input_file.bam

[dkolbe]

Regarding the Picard errors - I think it relates to the (mac) version of my java (which otherwise is up to date):

Exception in thread "main" java.lang.NoClassDefFoundError: jvm-args
[...]

You're using this as a template, I'm guessing:
java jvm-args -jar PicardCommand.jar OPTION1=value1 OPTION2=value2...

Like the command and the options, jvm-args needs to be substituted for actual java arguments. A typical example is -Xmx2g (specifying 2G of memory allocated for the run)