Using GATK HapCaller/CombinesGVCFs/GenotypeGVCFs as a java library. Any drawback ?

[lindenb]

Hi all,

That is not an issue but a technical question about using the 'jar' of gatk as a java library.

I want to genotype a lot of BAM files in a small region of the genome in the same process of my pipeline.
The JVM is slow on startup and I think the initialization process takes more time than the calling itself.

So I wrote a small java program that use gatk as a library and invokes all the steps in the same JVM instance:

So Each call to gatk is basically the following code:

public class Gatk4ProxyImpl  extends org.broadinstitute.hellbender.Main implements Gatk4Proxy {
(...)
	public void execute(final List<String> argv) throws Exception {
		final String[] args = argv.toArray(new String[argv.size()]);
		LOG.info(getCommandLineName()+":Executing:  gatk "+ String.join(" ",argv));
		final CommandLineProgram program =
			this.setupConfigAndExtractProgram(args, 
				this.getPackageList(),
				this.getClassList(),
				this.getCommandLineName()
				);
	    final Object result = Main.runCommandLineProgram(program, args);

(...)

(full code at https://github.com/lindenb/jvarkit/blob/master/src/main/java/com/github/lindenb/jvarkit/gatk/Gatk4ProxyImpl.java )

my question is: The code works but is there any hidden drawback in using gatk such way ? For example there any chance that I'm missing any function that dispose some resources, that loads a huge resource for each call, etc....

thanks,

Pierre