Data Processing Pipeline

mike_boursnell

Can the DataProcessingPipeline (as used by Queue.jar) be used to process BAM files in the best recommended way ("Best: multi-sample realignment with known sites and recalibration")

Carneiro

No, the DPP as is will split all bam files per sample and process them independently. Just to reiterate, the DPP script is an example script. You are welcome to modify it to do all samples at once, using the DPP as a template.

Carneiro

Although the Data Processing Pipeline follows our best practices guidelines, it is provided as an example script for people to build their own. You risk using parameters that are not updated or best for your particular data set.

Geraldine_VdAuwera

What my esteemed colleague @Carneiro means to say, I believe, is that the DPP script performs the necessary steps described in the Best Practices, but you should not simply run it with default parameters. You should check that the way the steps are set up make sense for your dataset and above all make sure you adapt the parameters to the specificities of your data.

mike_boursnell

Thanks for the folloe up. :-) What I was trying to ask is this: The "Best" recommended processing of BAM files seem to involve processing multiple BAM files together (i.e. in parallel rather than in series). Can the DPP do this, and if so how do I set it up to do that?