Matching Trinity fusion transcripts to FusionInspector TSV output

3 views

Skip to first unread message

Pavlína Slavníková

unread,

Mar 10, 2026, 8:37:57 AMMar 10

to Trinity_CTAT_users

Hello,

I am using FusionInspector to combine and validate results from multiple fusion callers. I planned to use the Trinity-generated fusion transcript sequences (via the --include_Trinity option) for downstream fusion gene expression estimation.

However, I am having trouble determining which sequence in the FASTA file (finspector.mm2_trinity_GG.fusions.fasta) corresponds to which entry in the TSV output (finspector.FusionInspector.fusions.tsv). The column TrinGG_Fusion, which I expected to contain this information, is not present in either finspector.FusionInspector.fusions.tsv or finspector.FusionInspector.fusions.abridged.tsv.

Additionally, there are multiple isoforms for a single fusion and breakpoint, and some fusions listed in the TSV(including ones with strong support) are missing from the FASTA file.

Is there a reliable way to match the Trinity-derived fusion transcript sequences to the FusionInspector fusion calls?

Thanks in advance,
Pavlina

Brian Haas

unread,

Mar 10, 2026, 8:44:33 AMMar 10

to Pavlína Slavníková, Trinity_CTAT_users

Hi Pavlina,

When you run FusionInspector with the Trinity assembly option, it should indicate the supporting Trinity transcript accessions in the TrinGG_Fusion of the report tsv file. You should also find a finspector.mm2_trinity_GG.fusions.fasta output file that has the Trinity assemblies that support the fusion calls, and the accessions from the tsv should match that fasta file. In the fasta file, the header further includes the fusion name and breakpoint info.

If no Trinity assembly was found to support the fusion breakpoint, then you'll find just NA in that column of the tsv. The Trinity support has high specificity but fairly low sensitivity in that it requires sufficient read support for being able to assemble the fusion transcript - whereas the primary FusionInspector approach only needs to find a couple of reads that support it.

hope this helps,

Brian

--
You received this message because you are subscribed to the Google Groups "Trinity_CTAT_users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinity_ctat_us...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/trinity_ctat_users/fba904d6-22fd-4848-8b9d-dd45a5dd31ean%40googlegroups.com.