In the current implementation input BAM files are merged to the alignment files produced by the pipeline mapping step and then used for downstream processing. This approach can pose problems because:
- merging external BAMs is risky as there is no guarantee that the data has been processed in a consistent way
- sometime project data is kept for long-term storage in BAM format but we want fully reprocess the data with
grape-nf (e.g. re-mapping with a different genome/annotation)
A possible better approach could be based on params.steps and act accordingly:
- if the
mapping step is specified the BAM files should be re-processed
- if only downstream steps (e.g.
quantification) are specified the BAM files are used as-is. An error should be reported If BAM files can be merged
Alternatives might be considered as well...
In the current implementation input BAM files are merged to the alignment files produced by the pipeline
mappingstep and then used for downstream processing. This approach can pose problems because:grape-nf(e.g. re-mapping with a different genome/annotation)A possible better approach could be based on
params.stepsand act accordingly:mappingstep is specified the BAM files should be re-processedquantification) are specified the BAM files are used as-is. An error should be reported If BAM files can be mergedAlternatives might be considered as well...