Java NullPointerException when running virusbreakend.sh #480
Comments
No. I dropped support when I found that the performance difference was >10x when using cram inputs. |
|
Hi, great, thanks for the clarification and update. I'm just re-opening this as it seems like the issue wasn't related to gridsstools. I've compiled it from source and verified that it's being used (The log contains:
Let me know if I'm best to open as a new issue. |
|
That error is caused by the This seems like there was an error building your virusbreakenddb database. What do you get when you run the following? You should have 0 lines returned. That is, every viral reference in your database should also be in seqid2taxid.map. |
|
If you still have all the files created by To verify that the seqid2taxid.map matches the reference sequences in your kraken2 database (both commands should return nothing). |
|
Yes, it looks like there are some viral references that aren't in the seqid2taxid.map file. They're all in the is rebuilding the best bet or is there a faster way to add these 35 references? Thanks for the help! |
|
I think I rebuild is needed as kraken2 updates seqid2taxid.map file as part of it's build process. Technically speaking, you could edit You didn't happen to save the build output to a log file did you? |
|
Hi Daniel, I tried building the reference files twice, but each time I ran into the same error described in this issue. In both cases the underlying issue was that kraken2-build excluded some 15,000 sequences from the I resolved it by removing all of the unmapped sequences from the file I looked up some of the sequences in the NCBI nucleotide database. Almost all of them seem to be fairly recent SARS-CoV2 virus genomes (less than two weeks old). Could it be that kraken2 fails to map these sequences, because their annotation is not complete yet? Maybe, the build would succeed a few weeks from now when the databases have been updated fully. If this is a frequent issue, it may be a good idea to implement a step in the script Regards, |
|
I have been having much trouble building or downloading a virusbreakend database. If you can resolve either issue I can move forward. INFO 2023-05-17 10:28:10 IdentifyViralTaxa Loading seqid2taxid.map from virusbreakendORGdb/seqid2taxid.map Please is there a suggestion to get around this error? 2.) I tried to download the minimal reference DB from web sites mentioned in the VIRUSBreakend_Readme.md but the downloads from both sites failed (download stopped) multiple times but always at the same point, at about 28 GB into the expected 34.5GB download. |
Hi, thanks for the tool! I've been able to run GRIDSS successfully and generate the kraken database with
virusbreakend-build.sh, but get an error from the virusbreakend.sh script when I run:bsub -M 20G "virusbreakend.sh sample.gridss.masked.vcf.gz --db virusbreakenddb --jar gridss-2.11.0-gridss-jar-with-dependencies.jar --reference GRCh38.d1.vd1.fa --output sample.gridss.masked.virus.vcf.gz sample_T.bamFirst, when using gridsstools, I received an error about not being able to load shared libraries: (
gridsstools: error while loading shared libraries: libbz2.so.1.0) but the script still continued and returned a Java error (see below). Since I'm on a shared cluster (without root), I wanted to avoid compiling gridsstools from source. I removed it from my path and edited the script not to exit when gridsstools isn't found (sidenote: the "2-3x time longer" error message is a bit confusing since it reads like a warning but still kills the program. Is running without gridsstools supported?).Both with and without gridsstools though, I received the following Java error at the end of the log file (edited to remove ids and full paths):
Let me know if I can provide any more info.
The text was updated successfully, but these errors were encountered: