Creating Annotations using Computational Results in Apollo. Broader questions regarding motif annotation.

Question

[sniderbl]

Hello everyone,

I have a bit of a complicated problem that I'm trying to solve, and I was wondering if anyone here on Biostars could possibly give me some helpful information.

TLDR:

#1 I have an error in the Apollo Desktop application that I could potentially use help with

or

#2 Potentially and Preferably would like to use Web Apollo 2.0. Does this program have the same functionality as the desktop application?

Expanded: Is web apollo capable of importing computational results (to overlay on an annotation) in a manner similar to that of the desktop application?

#3 During Web Apollo installation:

• "copy the war file at target/apollo-X.Y.war to your webapps folder as apollo.war or similar"*

-Which webapps folder is this line referring to, and where is it likely located?

#4 How can I combine multiple blast results from Blast+ Blastn (in XML format) into a single file? Does anyone have information on Game.xml or Chado.xml

I have been working for some time on an undergraduate thesis project in which I am trying to construct a pipeline program in Perl which can take a Fasta file containing a multiple sequences, blast them (using Standalone Blast+'s Blastn algorithm) against a series of databases containing known sequence motifs, and provide the blast results as a computational output to a program such as the Desktop version of Apollo (GMOD circa 2007), in order to allow an annotator to hand annotate the sequences from the results.

At present, I have successfully written a Perl script which can generate blast results using BioPerl calls to StandAloneBlast+, and the blast results appear to be accurate.

Now I am faced with trying to load these results into Apollo. The problem is that when I load up Apollo (desktop) and select "Computational Analysis Results" for "Choose Data Source" and use my program's output (a blast result from Blast+ in XML format (i.e., outformat => 5), and add my fasta sequence to "my query sequence file (FASTA format)" I get the following error:

Range.setFeatureType: can't accept feature type of null. Use SeqFeature.NO_TYPE or 'SeqFeatureI.No_TYPE' instead.

This makes it apparent that something is either off with my analysis settings or my files. However, I am confident at this point that my Blast output file and Query fast are properly formatted (willing to link these is someone needs to see them). So, my first question is, why am I getting this error? Is SeqFeature.NO_TYPE supposed to be declared within the blast result somewhere?

Because I was having these issues with Apollo, I was doing some digging and found out that the desktop version of Apollo is no longer supported by GMOD, and that a new version of apollo, Web Apollo, had replaced the desktop application. Thus I have set about installing Web Apollo to hopefully perform the same task, so question #2 is: Is web Apollo capable of importing computational results (to overlay on an annotation) in a manner similar to that of the desktop application? All of the documentation that I have seen so far implies to me that this is true, I would just like confirmation.

Currently I have successfully compiled (deployed) web apollo utilizing the documentation found here:

http://webapollo.readthedocs.org/en/latest/Apollo2Build/

However the documentation loses me here:

copy the war file at target/apollo-X.Y.war to your webapps folder as apollo.war or similar

(#3) Which webapps folder is this line referring to, and where is it likely located? I see a 'web-app' folder under the web apollo directory, but I don't believe that this is the folder which the documentation is referring to, as my searches for .WAR files bring me to documentation regarding TomCat.

Overall, my goal is to somehow be capable of combining all of the blast results for a given query sequence into a single file (similar to a Chado or Game XML file) such that when loading these files into an annotation program I only have to deal with 1 file instead of multiple blast results (1 for each database the query was blasted against). Then I would like to be able to hand generate an annotation with a program such as apollo using this "results" file. However, I am currently at a serious loss for how to go about doing this because I've had a really hard time finding documentation on Chado and Game XML formats.

Any help getting this up and running would be greatly appreciated.

Thanks in Advance

-Brandon

Answer 1

[cmdcolin]

#1 I have an error in the Apollo Desktop application that I could potentially use help with

As you have already suspected, desktop Apollo isn't supported at the time being, but Web Apollo is supported. I don't have any particular recommendations on the Apollo Desktop app as a result.

#2 Potentially and Preferably would like to use Web Apollo 2.0. Does this program have the same functionality as the desktop application?

WebApollo has much of the same functionality as the desktop version, but it is fully web-based and is built off of JBrowse. WebApollo can create user annotations using flexible drag and drops and edge matching of the annotations to the evidence, and can export the users annotations as peptide/cds FASTA format files and GFF format files, and it has been designed to retain all of the annotation features that desktop Apollo has.

#3 During Web Apollo installation:

"copy the war file at target/apollo-X.Y.war to your webapps folder as apollo.war or similar"

-Which webapps folder is this line referring to, and where is it likely located?

The webapps folder would generally be the webapps folder used for tomcat, which is a java based webapp server. On a linux machine, it can be /usr/share/tomcat7/webapps (redhat or similar) or /var/lib/tomcat7/webapps (ubuntu or similar), or you can also download it yourself. Also, WebApollo can automatically launch a temporary instance of tomcat (without you having to setup tomcat yourself) using "apollo run-local 8080" for example

#4 How can I combine multiple blast results from Blast+ Blastn (in XML format) into a single file? Does anyone have information on Game.xml or Chado.xml

I would only have recommendations for Web Apollo, but generally I would convert the blast results to GFF, and then load those GFFs into the genome browser via flatfile-to-json.pl or just with the File->Open menu. Multiple GFFs could probably be concatenated as well, as it is a simple tab delimited file format. I think GFF will represent the BLAST hits as "match" and "match_part" for subfeatures, and these are perfect for annotating in Web Apollo (the demo browser has a number of blast type tracks from similar species. note: login is demo:demo to view the annotation track).

Generally, I think this is a great project (and I think changing from desktop apollo to web apollo will be a good move). If you want any more specific help feel free to ask more or check out our mailing list

Disclaimer: I'm a Web Apollo developer

[sniderbl]

Thank you for the information. I think it will help me out a ton.

I'm about to try again to get Web Apollo running. I'm also going to try and read up more on WebApollo on its documentation pages. Would you mind if I contact you privately if I have future questions? I could really use a link to a Web Apollo Dev, as my project may involve me having to know Web Apollo pretty well. Also, I'm on a bit of a time crunch. My defense is approaching before summer's end.

Thanks again

[cmdcolin]

Sure I'd be happy to answer questions by email or chat.

[sniderbl]

Wow, thanks so much. So, I was having issues getting things set up, so now I'm working on setting up WebApollo via GMOD in a box.

[cmdcolin]

The GMOD in a box version is fairly old now, and it doesn't necessarily contain any bad bugs, but I can help with setting up the newer versions too.