[User Story] Speed up CADD indel annotation for WGS TO workflow

[mathiasbio]

Need

As a clinician I would like my WGS Tumor Only analyses to finish in a reasonable time. Currently there is a un-ideal filter in place which reduces the number of homozygote germline variants in the WGS Tumor Only analyses: #1345 but when this will be removed then the InDel CADD annotation step will take roughly 20hours to finish.

Suggested approach

To fix this problem we could probably add the gnomad population database annotation step to directly after quality filtration, then add a very relaxed gnomad population filtration. This can later be followed up by a more harsh filtration which can be whitelisted in the final step if they are pathogenic in clinvar (which is being worked on here: #1570)

Alternatively, it has been seen that the CADD rule may not have been properly configured to use multi-threading and has been using a default of 1 core when more may be possible. This is being tested here: #1602 and this is probably the preferred solution as we don't need to change the nature of our filtering strategy.

Considered alternatives

No response

Deviation

No response

System requirements assessed

• Yes, I have reviewed the system requirements

Requirements affected by this story

No response

Risk assessment needed

• Needed
• Not needed

Risk assessment

No response

SOUPs

No response

Can be closed when

No response

Blockers

No response

Anything else?

No response

[mathiasbio]

With 0.01 as gnomad_snv filter we see the following changes to the CADD indel annotation time:

	v17.1.0	this PR (with 0.01 gnomad filter)
case	cadd indel step time	cadd indel step time
WGS TO case 1	17:23:37	2:03:23
WGS TO case 2	6:03:02	1:43:02

However, I think these times may be very dependent on how busy the cluster is. Because in case 2 which was run over the weekend it took the noted time, 6 hours, but in a previous run it took over 17 hours. Khurram told me that the CADD script is running snakemake under the hood and maybe it then parallelises as much as possible on the node.

I just took a look at the CADD.sh script on the container, and apparently it's actually possible to set the cores with the -c argument. I'll try that in a separate PR and maybe setting these cores makes this change completely unnecessary...

[mathiasbio]

Discussed in refinement meeting: We're prioritising the CADD threads solution if it works.

[fevac]

if time, included in v18. Otherwise in next one