Skip to end of metadata
Go to start of metadata

Dear Sammeth,

I run flux-capacitor in the following options: flux-capacitor -i bamfile -a annotation_file -o ouptut --threads nthreads.

When I run it with an gff annotation file that have some transcripts that overlap each other, but not exons, such as I show in the following example with 3 transcripts or in the attached image dmel.overlap.jpg.

 

2L    selenoprofiles_exonerate    exon    4939284    4939788    .    +    .    transcript_id "FBtr0304686.74.homologue"; gene_id "FBgn0304686.74.homologue";
2L    selenoprofiles_exonerate    exon    4891851    4892080    .    +    .    transcript_id "FBtr0304686.74.homologue"; gene_id "FBgn0304686.74.homologue";
2L    selenoprofiles_genewise    exon    4893955    4893994    .    +    .    transcript_id "FBtr0077391.1.homologue"; gene_id "FBgn0077391.1.homologue";
2L    selenoprofiles_genewise    exon    4901402    4901917    .    +    .    transcript_id "FBtr0077391.1.homologue"; gene_id "FBgn0077391.1.homologue";
2L    selenoprofiles_genewise    exon    4902025    4902395    .    +    .    transcript_id "FBtr0077391.1.homologue"; gene_id "FBgn0077391.1.homologue";
2L    selenoprofiles_genewise    exon    4902461    4902593    .    +    .    transcript_id "FBtr0077391.1.homologue"; gene_id "FBgn0077391.1.homologue";
2L    selenoprofiles_genewise    exon    4902664    4902905    .    +    .    transcript_id "FBtr0077391.1.homologue"; gene_id "FBgn0077391.1.homologue";
2L    selenoprofiles_genewise    exon    4902974    4903150    .    +    .    transcript_id "FBtr0077391.1.homologue"; gene_id "FBgn0077391.1.homologue";
2L    selenoprofiles_genewise    exon    4937441    4937468    .    +    .    transcript_id "FBtr0077393.27.homologue"; gene_id "FBgn0077393.27.homologue";
2L    selenoprofiles_genewise    exon    4937534    4938182    .    +    .    transcript_id "FBtr0077393.27.homologue"; gene_id "FBgn0077393.27.homologue";
2L    selenoprofiles_genewise    exon    4938324    4938866    .    +    .    transcript_id "FBtr0077393.27.homologue"; gene_id "FBgn0077393.27.homologue";
2L    selenoprofiles_genewise    exon    4940964    4941363    .    +    .    transcript_id "FBtr0077393.27.homologue"; gene_id "FBgn0077393.27.homologue";

The ouptut results obtained for this example are the following:

2L    selenoprofiles_genewise    transcript    4893955    4903153    .    +    .    transcript_id "FBtr0077391.1.homologue"; locus_id "2L:4891851-4941366W"; gene_id "FBgn0077391.1.homologue"; reads NaN; length 1482; RPKM NaN

2L    selenoprofiles_genewise    transcript    4937441    4941366    .    +    .    transcript_id "FBtr0077393.27.homologue"; locus_id "2L:4891851-4941366W"; gene_id "FBgn0077393.27.homologue"; reads NaN; length 1623; RPKM NaN

2L    selenoprofiles_exonerate    transcript    4891851    4939788    .    +    .    transcript_id "FBtr0304686.74.homologue"; locus_id "2L:4891851-4941366W"; gene_id "FBgn0304686.74.homologue"; reads NaN; length 735; RPKM NaN

As it can be seen the reads and RPKM values are inconsistent. Some parts of the ouput of the execution are shown here. It can be seen that are some errors and warnings that are resposible for these NAN values.

PROFILE] Scanning the input and getting the attributes.

        profiling  OK (00:23:52)

        first round finished .. took 1432 sec.

        12325 single transcript loci

        377309560 mappings in file

        43276718 mappings fall in single transcript loci

        4615251872 mappings in annotation-mapped pairs

        81,121 min/max read length

[SOLVE] Deconvolving reads of overlapping transcripts.

decomposing  [WARN] Fraction inconsistency FBtr0077393.27.homologue     -2.304722170272954

...

[ERROR] Failed to set lp output to:
        null
java.lang.NullPointerException
        at java.io.FileOutputStream.<init>(FileOutputStream.java:206)
        at java.io.FileOutputStream.<init>(FileOutputStream.java:136)
        at barna.flux.capacitor.reconstruction.GraphLPsolver.run(GraphLPsolver.java:950)
        at barna.flux.capacitor.reconstruction.FluxCapacitor$LocusSolver.call(FluxCapacitor.java:1097)
        at barna.flux.capacitor.reconstruction.FluxCapacitor.explore(FluxCapacitor.java:2631)
        at barna.flux.capacitor.reconstruction.FluxCapacitor.call(FluxCapacitor.java:1886)
        at barna.flux.capacitor.reconstruction.FluxCapacitor.call(FluxCapacitor.java:74)
        at barna.commons.launcher.Flux.main(Flux.java:198)

        unusual normalization factor: 0.03384014257154664
[WARN] Unsolved system: 2L:4891851-4941363W
[WARN] Fraction inconsistency FBtr0302984.1.homologue   -2.313103458721854
Model name: FBtr0080341.1.homologue

 Is there a solution for this overlapping case?

Thank you.


Cheers,

Miqueldmel.overlap.jpg

 

  • No labels

2 Comments

  1. Hi Miquel,

    thank you for your efforts to describe the problem you observed in detail. The program behavior is apparently an error, I created a bug report. You can log in there with your wiki account, and "watch" the ticket to get automatically notified when me or some other developer updates on it.

    The example locus you describe should not be a problem, transcripts can of course overlap like that and it should not have any influence on the ability of the program to run. Currently it seems to me that some weird problem with creating a temporary file could be responsible for the error.

     I will try to reconstruct your problem and get back to you.

    Cheers,

    Micha

    1. Tried but failed to reconstruct the problem, would it be possible to isolate the reads from this region?

      see the Comment