order of taxonomic assignment, refDF flag

[bbroo...@gmail.com]

I was under the impression that the -refDB flag should assign taxonomy in the order in which you supply each database. So the below command should first do the assignment with silva and sequences that could not be assigned using silva, would then use GreenGenes. Correct? 

In looking at the log file, it seems the order is not preserved. Am I understanding the log file correctly (see pasted excerpt below)?

Thanks! 

From the LotuS_runlog.txt:

LotuS 1.502

=========================================================================

COMMAND

perl /home/bbrooks/software/lotus/lotus_pipeline/lotus.pl -i ../reads/ -m ../mapping_files/libs1-9_moreMetaData.txt -o lotus_len150_q20_lotus1.502_wHSpacers_BinaryFilterExp3Alph0.0005_sdm1.27_BCErr1Primer

Err5_lanes1-9_silva_GG_derep -s /home/bbrooks/software/lotus/lotus_pipeline/sdm_miSeq.txt -c /home/bbrooks/software/lotus/lotus_pipeline/lOTUs.cfg -threads 8 -refDB SLV.GG -highmem 1 -p miseq -derepMin 3:

10,10:3 -simBasedTaxo lambda

...............................

...............................

...............................

=========================================================================

          Merging OTU seed paired reads

=========================================================================

=========================================================================

          Removed 96 chimeric

          and 0 phiX contaminant

          OTUs (2278347 read counts) from abundance matrix,

          1549 OTUs remaining.

=========================================================================

=========================================================================

          Assigning taxonomy against reference using LAMBDA

          elapsed time: 21868 s

=========================================================================

Read greengenes taxonomy

2 / 1215 reads failed LCA assignments

Writing 336 of previous 1549 reads for next iteration.

=========================================================================

          Assigning taxonomy against reference using LAMBDA

          elapsed time: 26994 s

=========================================================================

Read SILVA taxonomy

2 / 2 reads failed LCA assignments, checked 336 reads.

Assigned greengenes SILVA Taxonomy to OTU's

Total of 1213 / 1215 reads have LCA assignments

=========================================================================

          Calculating Taxonomic Abundance Tables from greengenes SILVA assignments

          elapsed time: 27527 s

=========================================================================

Calculating higher abundance levels

Total reads in matrix: 70603181

TaxLvl  %AssignedReads  %AssignedTax

Phylum  83.418788170465 96

Class   83.3640002707527        90.7692307692308

Order   83.3180320869679        77.3109243697479

Family  81.9976340726064        71.6666666666667

Genus   37.8088771949241        61.1713665943601

Species 20.5244321782045        16.8582375478927

[Falk Hildebrand]

Hey,

the problem is in the command: "-refDB SLV.GG " it needs to be comma separated, so "-refDB SLV,GG ". In ver 1.504 I made the acceptance criteria harder, so that lotus will give an error next time the "." and "," are interchanged.

However, I noticed that the LCA C++ program was also not installed in your case. I strongly recommend using the C++ program for double DBs, as this is massively improved using the C code (for single DBs it doesn't matter that much, whether the C++ or the old Perl code is used).

I'm sorry, but it looks like you have to apply the now uploaded 1.504 update: 

1) run ./autoupdate.pl

2) run ./autoInstall.pl -forceUpdate

I hope this is the last time this incovenient update procedure has to be run, actually the 1.504 update is focused on the updated autoinstaller to avoid these inconveniences in the future.

hth,

Falk

[bbroo...@gmail.com]

Hi Falk,

My confusion came from your updated log post on 12th Feb 2016 in which you have HITdb.SLV listed. I realize this is an old version, just wanted to communicate the source of my confusion. But, your suggestion worked and assignment is now being performed in a predictable order. Thanks! 

I also ran the -forceUpdate and that too worked well. Thanks again for your quick reply. 

[Falk Hildebrand]

Hey,

Thanks, I changed it on the webpage.

best,
Falk