rep_set vs new_refseq fo chimera picking

[Samuel Major]

Hi, 

I think this is a simple question.

I picked open reference OTUs and it generated new_refseq.fna and rep_set.fna files. What is the difference between these two?

Also, I'm using vsearch --uchime_ref and --uchime_denovo to pick chimeras with these files and I'm recieving errors because there are "." characters in the files. Could someone advise me on the best uchime strategy to pick chimeras after I'm performed the OTU picking? 

A few of my commands and errors:

(qiime1) qiime@qiime-190-virtual-box:~/F16_Harold/F17_analysis/F16_16S/2_16S_SILVA_128_opref$ vsearch --uchime_denovo new_refseqs.fna --nonchimeras new_ref_denovo_nonchime.fa

vsearch v2.5.1_linux_x86_64, 10.3GB RAM, 1 cores

https://github.com/torognes/vsearch

Reading file new_refseqs.fna 7%  

Fatal error: illegal character '.' on line 26744 in FASTA file

(qiime1) qiime@qiime-190-virtual-box:~/F16_Harold/F17_analysis/F16_16S/2_16S_SILVA_128_opref$ vsearch --uchime_denovo rep_set.fna --nonchimeras new_ref_denovo_nonchime.fa

vsearch v2.5.1_linux_x86_64, 10.3GB RAM, 1 cores

https://github.com/torognes/vsearch

Reading file rep_set.fna 72%  

Fatal error: illegal character '.' on line 164996 in FASTA file

vsearch --uchime_ref new_refseq.fna --db ~/SILVA_128_QIIME_release/rep_set/rep_set_16S_only/97/97_otus_16S.fasta --nonchimeras rep_set_nonchime.fa

vsearch v2.5.1_linux_x86_64, 10.3GB RAM, 1 cores

...

233091219 nt in 166393 seqs, min 900, max 2961, avg 1401

Masking 100%  

Counting k-mers 100%  

Creating k-mer index 100%  

Detecting chimeras 7%  

Fatal error: illegal character '.' on line 26744 in FASTA file

Thanks

sam

[zech xu]

Hello,

The error message is pretty straightforward - do you have '.' in your fasta sequence?

[zech xu]

with regarding to chimera, you can find multiple related answers. For example: https://groups.google.com/forum/#!searchin/qiime-forum/chimera|sort:date/qiime-forum/QTSUx9pmS4A/oI2h6YTDDwAJ

[Samuel Major]

That's what I assumed. However, i'm not sure how a period would be introduced. I suppose there may have been an N that got substituted for an period. Unfortunately, based on your response and some reading, it does not seem like uchime can be used for post-OTU picked sequence chimera identification.