Weird problem of the blast algorithm with assign_taxonomy.py

[Chih]

Dear Qiime 1 users:

I got a weird result from using assign_taxonomy.py with the function -m blast.

Basically, the command worked and generated the log file as output file. The information in the log file is:

-------

lastTaxonAssigner parameters:

Application:blastn/megablast

Max E value:0.001

Min percent identity:90.0

id_to_taxonomy_filepath:/path/developer/sh_taxonomy_qiime_ver7_97_01.12.2017_dev.txt

reference_seqs_filepath:/path/developer/sh_refs_qiime_ver7_97_01.12.2017_dev.fasta

Number of sequences inspected: 75

Number with no blast hits: 75

Result path: /tmp/assign-tax3xRia6

---------

And nothing afterwards.

Thus, I think the incorporation of the qiime1 with blast might have some problems, but I cannot find what might be the reason for that.

However, if I used the same command (without indicating -m blast) by using the default algorithm, uclust, with exactly the same command. The log file is:

-------

UclustConsensusTaxonAssigner parameters:

id_to_taxonomy_filepath:/path/developer/sh_taxonomy_qiime_ver7_97_01.12.2017_dev.txt

id_to_taxonomy_fp:/path/developer/sh_taxonomy_qiime_ver7_97_01.12.2017_dev.txt

max_accepts:3

min_consensus_fraction:0.51

reference_sequences_fp:/path/developer/sh_refs_qiime_ver7_97_01.12.2017_dev.fasta

similarity:0.9

unassignable_label:Unassigned

Result path: /tmp/assign-taxPD_n9w

.uc file contents:

# uclust --input /path/data.fa --id 0.9 --rev --maxaccepts 3 --allhits --libonly --lib /path/developer/sh_refs_qiime_ver7_97_01.12.2017_dev.fasta --uc /tmp/UclustConsensusTaxonAssigner_DSfjiT.uc

# version=1.2.22

# Tab-separated fields:

# 1=Type, 2=ClusterNr, 3=SeqLength or ClusterSize, 4=PctId, 5=Strand, 6=QueryStart, 7=SeedStart, 8=Alignment, 9=QueryLabel, 10=TargetLabel

# Record types (field 1): L=LibSeed, S=NewSeed, H=Hit, R=Reject, D=LibCluster, C=NewCluster, N=NoHit

# For C and D types, PctId is average id with seed.

# QueryStart and SeedStart are zero-based relative to start of sequence.

# If minus strand, SeedStart is relative to reverse-complemented seed.

L 7293 716 * * * * * SH494153.07FU_HG935852_reps *

H 7293 257 93.4 + 0 0 10I211MI30MD15M449I OTU8.17739595034258e+28 SH494153.07FU_HG935852_reps

L 9750 658 * * * * * SH0354.07FU_AY046553_refs *

H 9750 257 92.2 + 0 0 37MI77MI74MD68M400I OTU8.17739595034258e+28 SH0354.07FU_AY046553_refs

L 19700 720 * * * * * SH006345.07FU_KF057516_reps *

H 19700 257 91.4 + 0 0 16I26MI11M3I46MI35M4I70MD54MI14M438I OTU8.17739595034258e+28 SH006345.07FU_KF057516_reps

L 5173 1137 * * * * * SH017876.07FU_JN890081_reps *

H 5173 182 97.3 + 0 0 6I87MI74MI21M947I OTU9.15302265991585e+28 SH017876.07FU_JN890081_reps

H 5173 181 97.2 + 0 0 6I11MI74MI76MI20M947I OTU9.90187917408672e+28 SH017876.07FU_JN890081_reps

H 5173 181 96.7 + 0 0 6I11MI74MI76MI20M947I OTU2.34003445922049e+28 SH017876.07FU_JN890081_reps
-------
(There are still many lines behind with similar formats)

Could anyone help me to figure out why they made such a difference?

Thank you.

Chih

[Lynn]

Hi, I am also facing problem like you. How do you solve it? I cannot install blast in my miniconda. So, I use default uclust. I am also analyzing for cbbM, functional gene for oxidase enzyme. If you know how to solve problem, please help me. thank u so much.

lynn

[Colin Brislawn]

Good afternoon,

I wonder if legacy blast (not blast+) is being installed correctly by conda. I know this has been an issue in the past and this would explain why uclust but not blast is working. 

Colin

[Lynn]

Dear Colin,

Thanks for your answer.

 I did not install legacy, only blast+. I don't know how to install it.

wget ftp://ftp.ncbi.nlm.nih.gov/blast/executables/LATEST/ncbi-blast-2.9.0+-x64-linux.tar.gz

tar zxvpf ncbi-blast-2.9.0+-x64-linux.tar.gz

export PATH=$PATH:$HOME/ncbi-blast-2.9.0+/bin

mkdir blastdb

export PATH=$PATH:$HOME/dell/miniconda3/wjs/ncbi-blast-2.9.0+/bin

rm ~/.ncbirc

export BLASTDB=~/home/wjs/blastdb

export BLASTMAT=~/home/wjs/blast-2.9.0/data 

I don't know what are these. But, I follow these commands and now, I can use blast in qiime. 

How to install legacy?

with best regards

lynn

[Colin Brislawn]

Hello Lynn,

You can install the old blast using conda: 

https://anaconda.org/bioconda/blast-legacy 

I'm not sure if you need 2.2.22 or 2.2.26 of blast-legacy. Try both and see which one Qiime likes.

Or, upgrade to Qiime 2!

Colin

[Lynn]

Thank you so much, Colin. I will install it.

best regards

lynn