fix(variant_to_vcf): increase batch job size

[project-defiant]

The gs://ot_orchestration/releases/24.10_freeze2/ produced VariantIndex dataset with ~ 2mln variants, which was aorund half less then expected - see variant to vcf batch job.

The cause of the credible_set failing the variant_to_vcf step was due to the Out Of Memory Exception thrown by spark session - see OOM Error

fix on VM heap size causing incomplete variant_index - variants_to_vcf was failing due to heap space.

After increasing the driver memory, the run was successful - variant_to_vcf

Things implemented:

• [x] Colocalisation step parameter update due to the fix implemented in gentropy https://github.com/opentargets/gentropy/pull/845
• [x] fix on VM heap size causing incomplete variant_index - variants_to_vcf was failing due to heap space
• [x] Removal of task that tests if release folder exists