DRL
commented
7 years ago Hi Jon,
I believe there is. There are actually two ways to deal with this:
- If you only care about the plots (quick and dirty) -> a)
- If you want to have the information in the BlobDB for subsequent filtering decisions -> b)
a) You can make a file of all contig_ids that have not been blasted (see below) and provide that file with the --catcolour.option to blobtools plot or covplot
contig_2,not_blasted
contig_10,not_blasted
contig_15,not_blasted
...b) Just generate a 'hits' file with the contig_ids you did not blast. You can specify as TaxID '0' which is 'root' in the NCBI taxonomy (and will result in all taxonomic ranks set to 'undef') and a score. E.g:
contig_2\t0\t1000
contig_10\t0\t1000
contig_15\t0\t1000
...You can then provide that file as an additional 'hits' file and when doing blobtools view -i blobDB --hits -r all you can recognise those by being undef at all ranks or based on the fact that they only got hits from that 'hits' file. However, if you want to plot them you still have to give it a catcolour file since otherwise they will be binned with other contigs that are annotated as 'undef' because of NCBI taxonomy (that happens if a taxon has not taxonomy at a given rank).
Let me know if this helps.
cheers,
dom
pinin4fjords
Hi,
For resource conservation we don't BLAST short contigs from our de novo assemblies. The consequence when using BAM files with reads mapped against a reference that does contain those contigs is Blob plots with grey clouds containing both contigs that weren't BLASTed and contigs that were BLASTed but produced no hits (and uninformative bars in the ReadCovPlot).
My proposal to deal with this is an additional parameter to specify the contigs we acutally BLASTed, producing plots with separate BLASTED and non-BLASTED no-hits categories. I'm happy to have a go at coding that if necessary. But is there a better way?
Thanks,
Jon