Cluster.split problem in second step of clustering

Re morning. I forgot to mention that I would use the lowest number of OTU. It is still high, but that is what you got.

Best of sucess.

Télécharger Outlook pour Android


Glad to hear that you solved your problem! I think I am having the same problem that you were having. Do you mind explaining which parameters did you change in the make.contigs command? In advance, thank you very much.


Glad you got something working. The risk of increasing the taxlevel argument is that there’s a possibility that a 3% OTU might get split across different taxa, which could artificially inflate the number of OTUs. You could also try increasing diffs in pre.cluster to 3 or 4.

I suspect your problem is a data quality issue. 2x150 nt reads will not fully overlap to denoise each other like 2x250 nt reads will.


Yes Pat, I understand the data quality is an issue.
I have another dataset where the reads are 2*300 nt. So will this provide a good overlap with SILVA database sequences?
The region used for PCR amplification for this dataset is is V3-V4. So what would be the start and end coordinates for pcr.seqs in this case? Please suggest.
The primers used are 338F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′).


For this I use PCR.seq on the silva aligned database and use the resulting file for alignment.

Best of success

Unfortunately, the 2x300 chemistry generates pretty poor 16S rRNA gene sequence quality. You can read more about this here:

For the coordinates of non V4 regions, I would encourage you follow this protocol: