Question about ground truth file requirement

[ksavhughes]

Hi Vitor Piro,

I would appreciate some clarification about the requirements for running metabench.

The readme says that
"It requires:
Reference sequences to build databases
Set of reads in fastq and ground truth in the bioboxes format"

I don't really understand what the ground truth in bioboxes is and how to obtain it or even where to put that in the config files. I see that *.profiling.bioboxes.gz and *.binning.bioboxes.gz are outputs for the classification function - is that what you are referring to?

Best,

Karissa

[pirovc]

Dear @ksavhughes, sorry for the delay answer.

The bioboxes refers to a file format defined in the CAMI challange. More information on the format for profiling and binning. If you have a custom set of reads you want to benchmark, you have to create those files to be able to obtain the full metabench report. Maybe check out the example files (1) (2). They are basically describing the origin of each read and the taxonomic abundance of the simulated sample.

The files you mentioned for the profiling and binning are the output of the tools that get converted the the bioboxes format for standardization.