Hi mothur users & developers,
I ran into a potential bug using the chimera.slayer in mothur v.1.19.3 for Mac.
mothur > chimera.slayer(fasta=1.TCA.454Reads.trim.chop.unique.good.filter.fasta, template=silva.gold.filter.fasta, processors=24)
Checking sequences from 1.TCA.454Reads.trim.chop.unique.good.filter.fasta …
Reading sequences from silva.gold.filter.fasta…Done.
Only reporting sequence supported by 90% of bootstrapped results.
[WARNING]: mothur found 0 potential parents, so we are not able to check G2S86LE01AX9Y0. This could be due to formatdb.exe not being setup properly, please check formatdb.log for errors.
[WARNING]: mothur found 0 potential parents, so we are not able to check G2S86LE01A8NHP. This could be due to formatdb.exe not being setup properly, please check formatdb.log for errors.
[WARNING]: mothur found 0 potential parents, so we are not able to check G2S86LE01BS5S5. This could be due to formatdb.exe not being setup properly, please check formatdb.log for errors.
This warning is given for most (not all) sequences. Eventually this resulted in a memory overflow ( Even with a small data set).
I noticed that this problem does not occur using v1.19.0 (haven’t seen v1.19.1 and 19.3 available for download).
What are the main differences between the subsequent versions? Can I keep using v1.19.0 for the time being :?:
Best wishes,
Guus
I encountered the similar problem…
mothur v1.19.3 Centos
mothur > chimera.slayer(fasta=Tongue26.unique.good.filter.unique.fasta, name=Tongue26.unique.good.filter.names, reference=self)
Checking sequences from /data1/baijiang/Run4_sample26_dup1/Tongue26.unique.good.filter.unique.fasta …
Sorting fastafile according to abundance…Done.
Reading sequences from /data1/baijiang/Run4_sample26_dup1/Tongue26.unique.good.filter.unique.fasta…Done.
Only reporting sequence supported by 90% of bootstrapped results.
[WARNING]: mothur found 0 potential parents, so we are not able to check s1a2. This could be due to formatdb.exe not being setup properly, please check formatdb.log for errors.
[WARNING]: mothur found 0 potential parents, so we are not able to check s1c2956805. This could be due to formatdb.exe not being setup properly, please check formatdb.log for errors.
[WARNING]: mothur found 0 potential parents, so we are not able to check s6a2875752. This could be due to formatdb.exe not being setup properly, please check formatdb.log for errors.
[WARNING]: mothur found 0 potential parents, so we are not able to check s1a2656491. This could be due to formatdb.exe not being setup properly, please check formatdb.log for errors.
[WARNING]: mothur found 0 potential parents, so we are not able to check s3a1612778. This could be due to formatdb.exe not being setup properly, please check formatdb.log for errors.
[WARNING]: mothur found 0 potential parents, so we are not able to check s1a5. This could be due to formatdb.exe not being setup properly, please check formatdb.log for errors.
…
To guus:
The only difference between 1.19.1 and 1.19.3 is the addition of the warning, and we included blast version 2.2.25 instead of 2.2.16. We added the warning to catch the case where you have the blast exes in the right spot, but have not set things up properly, so although blast runs it doesn’t return any results. Since you are getting results on some sequences, it seems you have blast setup properly. To be more clear the warning should probably be: [WARNING]: megablast returned 0 potential parents, so we are not able to check G2S86LE01AX9Y0. This could be due to formatdb.exe not being setup properly, please check formatdb.log for errors.
To baijiang:
Do ALL the sequences get the warning, or just most like guus?
This should just be a nuisance error message for those of you on os x that doesn’t mean anything. We’ve been having problems with the blast stuff for windows users and thought we had a way of fixing that problem and it turned out that we created another annoyance. We’ll have a work around for it shortly. Basically what is happening is that you are using reference=self which uses your more abundant sequences as references. So for the 3rd most abundant sequence it is looking to see if sequences 1 and 2 are the parents. If the blast comes back negative, mothur nags you.
I don’t know if the following is a bug, or if I just am not interpreting my chimera.slayer output correctly. I’m running a freshly downloaded Mothur v.1.19.4 on my Mac 10.6.7. Chimera slayer runs with no error messages, and I’m checking 524 900-bp Sanger reads against the Silva Archaea sequences that score >90 on Pintail and Sequence quality. In the output file, I get 58 chimeras and 253 non-chimeras, which is fine. The problem is that, for the remaining 213 sequences, there doesn’t seem to be any Chimera Slayer output at all. It just says “no” for each of them, and nothing else. I’ve spot-checked a couple of these with Pintail, and they weren’t chimeric, and there doesn’t appear to be anything anomalous (weird characters, or strange length).
Is “no” the response you get when both potential parents are the same? That would be nice, but somehow, I doubt the answer is that simple. I’ve been through the Haas et al. 2011 paper and the Broad Institute website, but I can’t find any explanation for this “no” thing. Does anyone else get this output?
Hi Karen.
There are 4 basic parts to the chimera.slayer algorithm:
Use blast to select closest potential parents.
Process the potential parents to find the best possible parents and breakpoints.
Realign the query sequence to the best parents
Process the query using bootstrapping
When the results say just “no”, it means that when mothur was running step 2 the sequence was determined to be not chimeric. In cases like this the Broad Institute outputs “ChimeraSlayer seqName NULL NULL -1 -1 -1 -1 -1 -1 NO”. I hope this helps.
The nuisance error message mentioned in my previous post ([WARNING]: mothur found 0 potential parents, so we are not able to…) does not appear running mothur v1.19.4.
Thanks a lot guys!
Guus
