Hello there,

I find that email/twitter is usually (unfortunatelly) more effective than tickets in repositories ;-) 

Anyway, if the problem persists try this solution I had discussed with Nabil some time ago and it worked for me:

1) Concatenate the query contigs (if they are multfasta files) . A way to do that in *nix systems is:

(echo ">concantenated"; grep -i '^[ACGTURYKMSWBDHVNX\-]*$' $INPUT_FILE | tr -d '\n' | fold -w 60; echo) > $OUTPUT_FILE

(substitute for your desired filenames or define the $INPUT_FILE and $OUTPUT_FILE; credit goes to https://gist.github.com/IgorFobia/6078848)

2) if you ran a previous analysis locate and remove the scratch folder on the output dir.

Tell me if it worked.

Cheers

This can be because of multiple reasons. e.g.

(1) Default low complexity region filter in BLAST might reduce the hits.

(2) There are no hits having more than 50% similarity. 

(3) one simply forgot or edited custom feature addition for visualization. 

(4) Technical problem within BRIG.

to dissect which one is the cause, you can run BLAST against one such genome and analyze the results. 

please add some more information. Which blast program, size of the genomes etc.

Thanks for your reply. I am now playing with just one of the genomes that fails. It does create ring if I use another genome as a reference instead of multi-fasta file, so there shouldnt be the problem with the file itself.

I checked all your points and none of them seems to be the reason of the failure. I have added BRIG working folder, if that could help to see the problem. Maybe you could suggest some other program which can be used instead of BRIG for the same purpose? 

BRIG working folder seems okay. However, It appears to me that you are using genbank format for query genomes. if it is so, why not try with simple fasta format. If you are using linux platform try dos2unix on all input files. 

Alternate tools would be CGview and Circos but for these you need to do a lot of file processing.

if possible attach one such genome input file as you answer. Multi-fasta is already in the BRIG working folder. 

Thanks for your suggestions, I will try them if we will not find solution for BRIG :) I enclosed one genome that fails either in .gbk or .fa format (Gal10) and one that always works well (Gal3). Thanks a lot for taking the time to help me!

Egle, did you ever find a solution? I am having the same problem. When I run BLAST directly I see that there are many high-identity hits, and there is a good tab output file from BLAST run with BRIG, but for some reason there are no features specified in the xml file and therefore no rings printed.

My files are protein fasta files, but I did choose blastp in the options in BRIG

I couldn't find a solution that time so I left it for a while. I also had many high-ID hits from BLAST, and .xml file seemed to be fine too (see the enclosed .zip files above). I am going to try to make it work again. I'll let you know if I get there...

I am having the same problem. In my case, the number of query genome is also 33. After I got the incorrect plot, I deleted the .xml file, modify the name of the output image, switch the format of the output image twice. Lastly I got the correct map. However, I don't think I find a solution. It seems sometimes the .xml file can't be generated with all features. I'm wondering if there is a stable solution. Maybe we need command-line based mode to run Brig?

Hello Egle

There was a compatibility problem of one of the Java versions that failed to create some of the BRIG rings.  I suggest that if you still have the problem update to the new java version and try again. If not contact Nabil , the creator of BRIG, and he will help you out.

Thanks João. I have updated Java but it didn't solve the problem. It seems there is something wrong with my files. We managed BRIG to work properly on one of our older computers with another sample set, but it still fails with mine when multi-fasta file is used as a reference. When I use ref genome and same query genomes it works fine. I hope Nabil will be able to find out what is wrong.

Hello Egle.

If you have a dataset that you can share with him that is not working and he can use it for testing he will certainly help out. Contact him via twitter @happy_khan.

Hey guys, just tried using brig today and I'm facing similar issues. I'm working with fasta files though and my reference is full genome not contigs. But I do have the issues that it is randomly not print rings. I say randomly because with the same strain and same reference, sometime it prints, sometimes not. It never print all the rings in my multiple stain plot (10 strains maximum) and I know they are similar enough to be plotted. I dropped a review on the sourceforge page but I highly doubt we're gonna get a solution from the author, as the last update of the tool was 2013, and the last time he replied to a ticket is 2011... It might be an issue of compatibility with the java version like mentioned previously since the software is from 2013 and I guess most of us have the latest java...

as an alternative, there is GenomeRing http://it.informatik.uni-tuebingen.de/?page_id=185 though it seems quite different but I didn't try it yet. 

Hello there,

I find that email/twitter is usually (unfortunatelly) more effective than tickets in repositories ;-) 

Anyway, if the problem persists try this solution I had discussed with Nabil some time ago and it worked for me:

1) Concatenate the query contigs (if they are multfasta files) . A way to do that in *nix systems is:

(echo ">concantenated"; grep -i '^[ACGTURYKMSWBDHVNX\-]*$' $INPUT_FILE | tr -d '\n' | fold -w 60; echo) > $OUTPUT_FILE

(substitute for your desired filenames or define the $INPUT_FILE and $OUTPUT_FILE; credit goes to https://gist.github.com/IgorFobia/6078848)

2) if you ran a previous analysis locate and remove the scratch folder on the output dir.

Tell me if it worked.

Cheers

I had similar results and I solved the problem by enter into the tab "Add custom features" and select multifasta in the tab "Input data". Then you "add" and "close". I got all the rings drawn.

Regards,

Alfonso

I also had similar problems. I tested BRIG_examples according to the manual. When the multi-fasta file "EHEC_vir.fna" was used as reference, only O126 genome couldn't be created. What's more strange, I could get the correct result when i pressed "submit" again. And the results were unstable. Sometimes O126 genome existed, sometimes not.

The O126 genome was just the best-match

I have solved the problem. We should install Java 1.6 (an old version) rather than update Java. BRIG was  written in Java 1.6. I have tried successfully. All the ring can be drawn.

Opening the program with the command line using "java -jar BRIG.jar" in the BRIG-0.95-dist directory seemed to fix the problem for me. Before I was seeing the same issue everyone is describing, but was opening the program with the Mac OS "open" command (i.e., "open BRIG.jar", which I believe is the equivalent of double clicking on the file in the finder window. Anyway, the java I'm running (from /usr/bin) is version "1.6.0_65" (you can check by typing "java -version" in your terminal/command prompt.

@Ian J. Miller I tried your method in Java "1.8.0_101". It also work incorrectly. For the ring which couldn't be drawn, "java.lang.IllegalArgumentException" was shown. So I think we should use Java 1.6

To be clear, I was using Java 1.6 when it worked for me.

Hello Egle, I had the same problem. The solution was uninstall the latest java version and install Java 1.6. In the archive of Oracle you can find this old versions. http://www.oracle.com/technetwork/java/javase/downloads/java-archive-downloads-javase6-419409.html

Hello, I have similar problems, so I will try the older Java version as suggested. In addition I have another problem with BRIG. I used a published WGS as a reference to compare with 2 WGS from strains that we have been sequencing. These 2 strains contain a large genomic element not present in the reference. Therefore, I would expect a gap in the ring of the reference sequence at that position, but it does not appear. Does anybody has a clue why?

HI Jannetta

BRIG just shows what you have in the reference. If your strains have novel elements they never appear in the BRIG representation. To illustrate that you have to use one of such strains as reference.

Cheers

I had the same problem... really worked after installing java 1.6. Thanks

Same here, using java 1.6 is inevitable. Also, on Ubuntu 16.04 I have to start the jar with administrator rights.

>sudo /pathtoJava1.6/bin/java -jar BRIG.jar

I didn't want to revert to Java 1.6. I discovered that removing all the 'Ring*' files under the scratch directory worked for me. But I have to do it every run.

I had the same problem in Ubuntu 18.04 and downloading and extracting the java version in the jdk-6u45-linux-x64.bin file downloaded from oracle did the trick. Regards.

For me it worked when I used the gbk file. I did not change my version of java