I am running chimera.uchime and getting the following:
mothur > chimera.uchime(fasta=xaa.fasta, count=stability.trim.contigs.good.unique.good.filter.unique.precluster.count_table, dereplicate=t,processors=2, reference=silva.v4.fasta)
Using 2 processors.
You have provided a countfile and the reference parameter is not set to self. I am not sure what reference you are trying to use, aborting.
[ERROR]: did not complete chimera.uchime.
Apparently mothur doesn’t like having the reference inserted externally for some reason. It does “work” (i.e. not giving errors) when using reference=self, but I’m not sure it’s actually doing anything meaningful. A very quick googling found out it may be some sort of a known bug. Do you have any ideas what’s going on and how I can work-around it?
chimera.uchime runs in two ways. the first is without a reference - it uses your own data as the reference by treating the abundant sequences as good and the rarer sequences as potentially suspect. this is why a names or count file is needed. the second is in reference mode where you give it a reference of non-chimeric sequences and it compares each sequence in your dataset against the reference.
we strongly encourage the use of the first approach in our SOPs and elsewhere. the problem with the second approach is that there are no good databases. the only sequences you can trust are those from cultured isolates and most phyla do not even have a single cultured representative. So just do this as it recommends in the SOP:
chimera.uchime(fasta=xaa.fasta, count=stability.trim.contigs.good.unique.good.filter.unique.precluster.count_table, dereplicate=t,processors=2)
Does it mean that first approach is in de novo mode?