can't tell if chimeras have been removed

[Laura]

Hi all

Have just successfully (I hope) run pick_open_reverence_otus.py (on my split_libraries .fna file) using the usearch61 option rather than uclust default. Now, having looked at the overall log file, I see no mention of chimeras at all. I just wanted some reassurance that running this

pick_open_reference_otus.py –i ~/mineral/seqs.fna –o ~/mineral/open_ref_otus_usearch –m usearch61

has actually done what I wanted, i.e. identified and removed chimeras, generated OTUs and assigned taxonomy. It seems a bit too good to be true and as there is no mention of chimeras.........is there someplace in any of the output files where the number of chimeras detected is recorded?

If someone can please reassure me that I'm correct, or alternatively show me where any issues are.....thanks in advance.

Laura

[Daniel McDonald]

Hi Laura,

I do not believe the open reference pipeline performs chimera checking by default. If you are concerned about chimeras within your study, or the questions you're asking, then I believe you should be able to add:

pick_otus:suppress_reference_chimera_detection False

...to a parameters file, and pass that into the pick_open_reference_otus.py script using -p

Best,

Daniel

[Laura]

Hi Daniel

Thank you very much for your reply. I have modified it so that I've included a parameters file containing the line that you've suggested. Having run it just now there appears no mention still of chimeras in any of the logs. This I find very strange indeed so I imagine that no chimeras have been removed. I've found a thread on the forum on the topic from 2013 (https://groups.google.com/forum/#!topic/qiime-forum/JsPgAYifM1A) and there it appears that the chimera checking isn't done within the pick_open_reference_otus.py command either, with the chimera checking done beforehand (actually before pick_otus). The paper I'm using as a guideline is that of Navas-Molina et al from July 2015, which states "uchime is the fastest method for detecting chimeric sequences and it is executed by default if the usearch method is selected for picking OTUs". So contradictory information.

So I guess my question still stands: on the webpage for the pick_open_reference_otus.py script there is absolutely no mention of chimeras either, which is leading me to believe that this script does not have options for chimera removal within it.

As for me, I've now checked for and removed chimeras with identify_chimeric_seqs.py and filter_fasta.py and will proceed from there, in the confidence that the chimeras have been removed.

All the best,

Laura

[Daniel McDonald]

Hi Laura,

I apologize for the challenges! I'll reach out to someone who has more detailed knowledge of that workflow.

Best,

Daniel

[Laura]

Hi Daniel

Not your fault at all, no apologies required. If I've realised anything in the short time I've been working with qiime it's that, as there seems to be many ways of accomplishing what you want, usually with very subtle differences in the output, it makes deciphering the workflow intricacies very difficult. Probably not helped in any way by subtle changes to different versions of qiime either!

All the best

Laura

[Jose Antonio Navas Molina]

Hi Laura,

Can you send me/post your log file?

Thanks!

[Jose Antonio Navas Molina]

Hi Laura,

I double-checked with other developers and we do apologize that the documentation can be a bit misleading. Qiime supports 2 different versions of usearch: 5.2 and 6.1. When running pick_open_reference_otus.py with the parameter "-m usearch" it is using the version 5.2 and when using "-m usearch61" is it using the version 6.1. Chimera checking is executed by default when using version 5.2, and that's why Navas-Molina July 2015 reads as "uchime is the fastest method for detecting chimeric sequences and it is executed by default if the usearch method is selected for picking OTUs". That is, when running with "-m usearch" it will execute chimera removal, but when running "-m usearch61" it will not. This should've been more clear in that manuscript, sorry for the inconvenience.

The chimera detection was separated from the otu picking in version 6.1 in order to give more freedom to the user to execute each step independently. However, you can still run chimera detection with usearch61 using identify_chimera_detection.py.

Hope this helps!

[Laura]

Hi Jose

Many thanks for your reply and apologies for the delay in replying, I've only just seen it. Your suggestion above is in fact what I actually did in the end. I assumed that the chimera checking failure was a version issue, so no problem. I just wasn't too trusting of my understanding so wanted some expert feedback. I really like your paper by the way, most helpful for a novice of qiime like myself and it explains the steps and theory very clearly.

All the best

Laura.

[Rune Grønseth]

I have a somehow related question, which probably will reveal my ignorance, but anyhow:

1. If you have a basic installation of macqiime and want to use greengages for open reference oto picking afterwards you pass 

identify_chimeric_seqs.py -i seqs.fna -m usearch61 -o usearch_checked_chimeras/ -r /macqiime/anaconda/lib/python2.7/site-packages/qiime_default_reference/gg_13_8_otus/rep_set/97_otus.fasta

then you filter the chimeras out by 

2. filter_fasta.py -f seqs.fna -o seqs_chimeras_filtered.fna -s usearch_checked_chimeras/chimeras.txt -n

Would the next step be to pass pick_open_reference_otus.py -i seqs_chimeras_filtered.fna -o pickedotus/ 

In the tutorial  ( http://qiime.org/tutorials/chimera_checking.html#usearch-6-1) it says to run pick_otus.py -m usearch61 -i seqs_chimeras_filtered.fna -o usearch61_picked_otus/

Thankful for any help (I'm also in your 2015 paper fan club)!

Best wishes rom Rune 

[TonyWalters]

Hello Rune,

You are correct-the usearch61 chimera-filtered file can be used with the pick_open_reference_otus.py script (pick_open_reference_otus.py is a workflow script that starts with the pick_otus.py script, for clarification regarding the above chimera filtering instruction page).

Best regards,

Tony