Hello community!
I am following MiSeq SOP
I am doing a run and while i am waiting for my alignment to be complete, I am rechecking and further investigating on how does mothur works.
Wo I was wondering…
After my make.contigs, there is nothing in my scrap file.
The sequencing center said I had a good run, but I find this supcicious.
yes, my reads do not fully overlaps… i should have check the forum before doing the run as recommanded by a collaborator but what is done is done
anyway, as i do understant it, it means that when there is no overlap, the quality of the basepair must be at least 25 to be kept and not to be scraped.
Is this enough? Can/should it be increase?
Many thanks for Mothur and good luck with the exploding use of MiSeq and the subsequent increase of questions!
So overnight sometimes helps to clarify matters
From what I get of this command make.contigs.
Overlap reads: there is a chance of correcting a base pair based on the difference of q scores if 2 aligned base pair do not match. If quality is too low or too much strong disagreement (forward read say A but reverse say G but both have a quality) the base is changed for N. If there is a gap, (forward read say “A” but reverse say gap “-”, the base should have a q over 25 to be called ok
(In the SOP I red 25 but in the command the parameter says: insert
The insert parameter allows you to set a quality scores threshold. When we are merging the overlapping regions, in the case where we are trying to decide whether to keep a base or remove it because the base is compared to a gap in the other fragment, if the base has a quality score below the threshold we eliminate it. Default=20)
If not ok, it is change to N
Down the road of the SOP, in the screen.seqs command, these sequences with ambiguous base call are removed.
For the non-overlap, from what I read, is that quality is not check at all? So if quality is low, there will be a lot of different base pair and this will create numerous wrong version of an ok read, thus make more unique reads that should actually be the same read, inflating the number of unique sequences.
So how does scrap is determined in this operation?
Is a solution to use trimoverlap? In my case it is not so bas since I have a 2x250 run for 300pb v4 fragment but it kind of defeat the purpose of the 300 pb original design…
The scrap file is usually empty (we should be deleting that file, this is a bug). If your reads do not overlap then we will make a contig out of them anyway. It is up to you in subsequent steps to remove those sequences in screen.seqs by setting a minimum or maximum length, minimum overlap length, and removing any sequence with an N. I would also strongly discourage sequencing regions where the reads do not fully overlap:
http://blog.mothur.org/2014/09/11/Why-such-a-large-distance-matrix/
Pat
Thanks you for your response. I am always amaze how dedicate you are about answering the community.
Is there a Canadian Workshop coming? I am in the province of Québec and having a Mothur workshop would most probably interest a lot of my colleagues as well as myself!
Unfortunately I went for what we were told to follow by a collaborator and now we have to do what we can with the run. I read the posts and blogs after I received my run raw files. We will most probably switch 16S primers if we ran into too much trouble in the end.
I was wondering if setting trimoverlap to “T” would change the results in the end? Has anyone tried this? I will try it but now the computer is dedicated to something else and if anyone can give me insight on this it would save myself a couple of analysis days.
Many thanks again.
I was wondering if setting trimoverlap to “T” would change the results in the end? Has anyone tried this? I will try it but now the computer is dedicated to something else and if anyone can give me insight on this it would save myself a couple of analysis days.
This would trim the contigs to the region that overlaps. I don’t think it will do much to help in your case and you’ll end up with mostly shrapnel
Is there a Canadian Workshop coming? I am in the province of Québec and having a Mothur workshop would most probably interest a lot of my colleagues as well as myself!
I frequently joke that anything east of Ann Arbor (e.g. Detroit) is Canada 
Unfortunately, I’m pretty limited in my travel. The next mothur workshop will be in late August.
Pat
Alright and Thanks again!
If you want to make a grand Canadian debut, just contact me, It would be my pleasure to organize it in the best of conditions possible.
Cheers