Tophat Fusion Post Processing filtering out confirmed positive fusion

[mknut]

I have been investigating a set of RNA-seq samples with known, confirmed fusions with Tophat-Fusion, in order to check efficacy of the tool and our pipeline. 4/5 fusions in different samples have been detected successfully. The fifth one was initially detected by Tophat (i.e. it is listed in fusions.out), but it was filtered out by post processing. I am trying to figure out why.

I have checked specificity of the fusion site manually with BLAST and the highest score I found was 110(length+identity percent), below the cut-off of 160.

The fusion is well anchored into both sites, and it also has sufficient number of supporting reads, pairs, and spanning pairs, as seen in its details, taken from fusions.out:

chr5-chr12      149510224       12006494        fr      38      13      19      0       86      79      0.546399        @       11 25 38 52 66  @       TTCCCCACTGTCAGGGTGGCTCTCACTTAGCTCCAGCACTCGGACAGGGA CTGCATGGAGAGAGCACTGAGTTAGGAGGCGGGAGGGTCAGGACAGTTAA   @       CCTGGATTTGTCTAAACCTCAGGCAAGAAAAGAGAAACCTCTTCCAGTAC CTTCTTCATGGTTCTGATGCAGTATGACCTCCGGCTGTGTGTGTATAGAG   @       38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 37 37 37 37 36 36 36 36 35 35 34 34 34 34 34 34 32 32 31 31 31 31 30 28 25 24 23 22 20   @       38 38 38 38 38 38 38 38 38 38 38 38 38 38 37 37 37 37 37 37 37 36 35 35 35 35 35 35 34 33 33 33 32 30 30 27 26 24 23 23 23 22 22 21 21 21 21 20 20 19   @       11:7 32:2 22:43 82:2 43:70 29:97 72:69 114:97 114:97 727:31 880:29 4046:105 5235:105

I had a look at the source code of tophat-fusion-post, and unfortunately I cannot find any indication as to why this specific fusion was filtered out.
Its check log does not exist in tophatfusion_out/check/

Commands used:
Tophat:

tophat -o tophat_sample_282 -p 8 --fusion-search --keep-fasta-order --bowtie1 --no-coverage-search -r 0 --mate-std-dev 80 --max-intron-length 100000 --fusion-min-dist 100000 --fusion-anchor-length 13 --fusion-ignore-chromosomes chrM /scratch/EXOME_DATA/RNASEQ/index_bowtie1/hg19 /scratch/EXOME_DATA/RNASEQ/positive_control_WTCHG/tophat/sample_282/WTCHG_cat_282_1.fastq /scratch/EXOME_DATA/RNASEQ/positive_control_WTCHG/tophat/sample_282/WTCHG_cat_282_2.fastq

Post processing:

tophat-fusion-post -p 8 --num-fusion-reads 1 --num-fusion-pairs 2 --num-fusion-both 5 /scratch/EXOME_DATA/RNASEQ/index_combined/hg19

Software versions:
Bowtie 0.12.8
Tophat 2.0.6
Blast 2.2.25
Samtools 0.1.18

I have emailed the creators of Tuxedo suite, and will post here if I get a reply.

Maybe there is something I am forgetting about? Some post processing condition that this specific fusion does not meet? Any help would be greatly appreciated.

[NKAkers]

Did you ever get an answer to this? I'm encountering something similar, I'd be interesting in hearing what you found.

[mknut]

Nothing at all, sorry.