search

Teichlab / tracer

TraCeR - reconstruction of T cell receptor sequences from single-cell RNAseq data
Other
121 stars 49 forks source link

igblastn and igblastp return an error #65

Open comrade1809 opened 6 years ago

comrade1809 commented 6 years ago

Hi,

I'm having an issue running igblastn. Every time I do, I get an error saying an annotation database in internal_data could not be found. I did follow the installation instructions. igblast is installed in "/tools/GNU/igblast/1.8.0" PATH is set to "/tools/GNU/igblast/1.8.0/bin" LD_LIBRARY_PATH is set to "/tools/GNU/igblast/1.8.0/lib" $IGDATA is set to "/tools/GNU/igblast/1.8.0/bin" I did download the internal_data folder from ticket 48 and placed the whole internal_data folder in the /bin directory as instructed. The error is still there.

Fully copy of the error is as follows:

Tuesday, March 13, 2018: 14:25:05       CMD: /.automount/tools/GNU/trinity/2.5.1/util/support_scripts/get_Trinity_gene_to_trans_map.pl /projects/Illumina/SolexaPipeline-0.2.2.6/W_eland_test/tracer/tracer-master/tracer-master/assemble_test2/cell_name/Trinity_output/Trinity_cell_name_TCR_B.Trinity.fasta > /projects/Illumina/SolexaPipeline-0.2.2.6/W_eland_test/tracer/tracer-master/tracer-master/assemble_test2/cell_name/Trinity_output/Trinity_cell_name_TCR_B.Trinity.fasta.gene_trans_map
shell-init: error retrieving current directory: getcwd: cannot access parent directories: No such file or directory

##Running IgBLAST##
BLAST query/options error: Germline annotation database mouse/mouse_TR_V could not be found in [internal_data] directory
Please refer to the BLAST+ user manual.
mstubb commented 6 years ago

Hi, have you also set the $IGDATA environment variable as described in issue #48?

Thanks,

Mike

mstubb commented 6 years ago

Oh sorry. Just seen that you did!

I’ll look into this tomorrow.

comrade1809 commented 6 years ago

Quick update: When I run 'igblastn' or 'igblastp' without any other options, I get an error stating BLAST Database error: No alias or index file found for nucleotide database [human_gl_V] in search path [/.automount/tools/GNU/igblast/1.8.0::] This is with the $IGDATA pointing to the bin directory as stated in issue #48.

mstubb commented 6 years ago

If you have the full internal_data directory installed under /tools/GNU/igblast/1.8.0/bin and you've set $IGDATA to be /tools/GNU/igblast/1.8.0/bin then I'm not totally sure what the problem is.

Could it be something to do with the fact that igblastn seems to think that it lives in /.automount/tools/GNU/igblast/1.8.0? What happens if you set $IGDATA to be /.automount/tools/GNU/igblast/1.8.0/bin?

PS - running igblastn without other options will always give that error message even if it's installed correctly. The best way to test will be to use tracer test -r where the -r switch will make it resume with previous assemblies so you don't waste time re-running Trinity.

Let me know how you get on...

comrade1809 commented 6 years ago

I did just change the IGDATA to what you suggested, but it did not help. I believe the .automount folder is a simple mount point for /tools so either that or /tools would have been correct in the eyes of Linux mounts.

After the change, I did run the application and got the following error.

[wbarnes@gencpu01 Illumina]$ /tools/GNU/igblast/1.8.0/bin/igblastn
BLAST query/options error: Germline annotation database human/human_V could not be found in [internal_data] directory
Please refer to the BLAST+ user manual.
mstubb commented 6 years ago

Please can you send me the output of the following commands:

ls /tools/GNU/igblast/1.8.0 ls /tools/GNU/igblast/1.8.0/bin ls /tools/GNU/igblast/1.8.0/bin/internal_data

Thanks,

Mike

comrade1809 commented 6 years ago

It's not pretty because of formatting but here you go:

[root@pyrite ~]# ls /tools/GNU/igblast/1.8.0 bin include internal_data internal_data_BACK lib optional_file

[root@pyrite ~]# ls /tools/GNU/igblast/1.8.0/bin blastdb_aliastool blast_formatter convert2blastmask gc_cli igblastp makembindex rpsblast seqdb_demo tblastn vdb_test blastdbcheck blastn csra_test_mt gene_info_reader internal_data makeprofiledb rpstblastn seqdb_perf tblastx wgs_test blastdbcmd blastp datatool graph_test legacy_blast.pl project_tree_builder run_with_lock snp_test test_pcre blastdbcp blastx deltablast igblastn makeblastdb psiblast seedtop sra_test update_blastdb.pl

[root@pyrite ~]# ls /tools/GNU/igblast/1.8.0/bin/internal_data human mouse rabbit rat rhesus_monkey

On Thu, Mar 22, 2018 at 6:35 AM Mike Stubbington notifications@github.com wrote:

Please can you send me the output of the following commands:

ls /tools/GNU/igblast/1.8.0 ls /tools/GNU/igblast/1.8.0/bin ls /tools/GNU/igblast/1.8.0/bin/internal_data

Thanks,

Mike

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Teichlab/tracer/issues/65#issuecomment-375307561, or mute the thread https://github.com/notifications/unsubscribe-auth/AZqh1P2ZjA_O1N7hqS38kmepvC606aIKks5tg6iEgaJpZM4Spk8u .

mstubb commented 6 years ago

Thanks. Please can you send me the path where you've installed Tracer and also the config file that you're using.

comrade1809 commented 6 years ago

Hi Mike,

Here's the path for Tracer: /tools/GNU/python/2.7.13/bin/tracer I don't know what you mean by the config file. I'm an admin that installs and fixes the applications our scientists use, so I'm not too familiar with how these applications run or what is required to make them run.

On Fri, Mar 23, 2018 at 4:25 AM Mike Stubbington notifications@github.com wrote:

Thanks. Please can you send me the path where you've installed Tracer and also the config file that you're using.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Teichlab/tracer/issues/65#issuecomment-375629163, or mute the thread https://github.com/notifications/unsubscribe-auth/AZqh1HaYl50dW9fycFpznMDgrvk4Ebwjks5thNuygaJpZM4Spk8u .

mstubb commented 6 years ago

Thanks for this. The config file details some settings for TraCeR and is described here: https://github.com/teichlab/tracer#setup.

It's found as follows:

TraCeR looks for the configuration file, in descending order of priority, from the following sources:

The -c option used at run time for any of the TraCeR modes The environmental variable TRACER_CONF, which can be set to point to a default location The default global location of ~/.tracerrc If all else fails, the provided example tracer.conf in the directory is used

ktpolanski commented 6 years ago

Hey, sorry to butt in, but does your server support Docker by any chance?

comrade1809 commented 6 years ago

@Mike I will check with the researcher if they are running any special config files. I was not able to find any in the Tracer base install directory located in site-packages of Python 2.7.13.

@Krzysztof, We can support Docker but due to the nature of the applications and research, it would be a short term solution. We have a CPU and GPU cluster that we use for processing big data, and in our experiments, Docker wasn't able to hold up in those environments. I know that NVIDIA came out with their own Docker support but we haven't had much success with it yet.

On Fri, Mar 23, 2018 at 9:01 AM Krzysztof Polanski notifications@github.com wrote:

Hey, sorry to butt in, but does your server support Docker by any chance?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Teichlab/tracer/issues/65#issuecomment-375714287, or mute the thread https://github.com/notifications/unsubscribe-auth/AZqh1ASgQcDyEHijqzDCkjhT5z_4kfaKks5thRxjgaJpZM4Spk8u .

comrade1809 commented 6 years ago

Unfortunately the researcher wasn't able to provide a config file or -c string as she was using igblastn as the application and not Tracer. I did check the users home dir and there is no .tracerrc file in their profile. There are also no TRACER_CONF variable set in the environment when loading the application.

On Fri, Mar 23, 2018 at 9:13 AM Aleksey Galipchak agalipchak@gmail.com wrote:

@Mike I will check with the researcher if they are running any special config files. I was not able to find any in the Tracer base install directory located in site-packages of Python 2.7.13.

@Krzysztof, We can support Docker but due to the nature of the applications and research, it would be a short term solution. We have a CPU and GPU cluster that we use for processing big data, and in our experiments, Docker wasn't able to hold up in those environments. I know that NVIDIA came out with their own Docker support but we haven't had much success with it yet.

On Fri, Mar 23, 2018 at 9:01 AM Krzysztof Polanski < notifications@github.com> wrote:

Hey, sorry to butt in, but does your server support Docker by any chance?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Teichlab/tracer/issues/65#issuecomment-375714287, or mute the thread https://github.com/notifications/unsubscribe-auth/AZqh1ASgQcDyEHijqzDCkjhT5z_4kfaKks5thRxjgaJpZM4Spk8u .