The .aggregated.metrics.json file is not always generated

[lukaas33]

Installation Type

Docker

pVACtools Version / Docker Image

griffithlab/pvactools:latest

Python Version

No response

Operating System

No response

Describe the bug

I have previously run the

How to reproduce this bug

I have previously ran the pvactools on a vcf file and got the .aggregated.metrics.json file in the output.

I then generated the vcf file from raw fastq files and ran the following command. I did get output but not the .aggregated.metrics.json file. Therefore I cannot visualise the results in pvacview.

`pvacseq run input.vcf samplename HLA-A*02:01,HLA-A*23:01,HLA-B*42:01,HLA-B*27:05,HLA-C*17:01,HLA-C*01:02 BigMHC_IM DeepImmuno pvacseq/ -e1 10 -e2 15 --iedb-install-directory /opt/iedb -t 8`

My main question is, what could cause the aggregate file not being generated. Is there a flag that I am missing?

Input files

Because of the sizes this is infeasible.

Log output

I unfortunately do not have this. I will rerun the analysis to get these logs later today.

Output files

MHC_Class_I.zip

[susannasiebert]

This behavior is by design. The metrics file is only created when at least one binding affinity algorithm is chosen, because the binding affinity data is core to what is displayed in pVACview. Immunogenicity algorithms, like BigMHC_IM and DeepImmuno, (as well as elution algorithms) only provide supplemental data to pVACview/the metrics file. We've tried to describe this behavior in the description of the metrics file in our documentation.

If you add at least one binding affinity algorithm to your run, you will see the file getting created.

[lukaas33]

Ah thank you, I will run it again.
Is it possible to display a warning in this case, or even an error?
Because the output does not seem very useful right now right?

[susannasiebert]

pVACseq would still generate a all_epitopes.tsv (no aggregated) and filtered.tsv file, which could be useful for some of our users. But adding a warning is a good idea.