Sebastian Uhrig

Awesome! Then I can finally close this issue. I will apply the fix to the dev branch. Hopefully, I will find some time soon to release a new version of...

Hi Silvia, I haven't run Arriba run on SMART-Seq2 data. I will give this a try and report back. Arriba automatically adjusts to the sequencing depth and has quite sensitive...

Hi Sogi, Do you have 10X Genomics scRNA-seq data or full-length transcripts? You're using Cell Ranger for alignment which suggests the former. The problem with 10X data is that the...

If it's a 5' library, things are slight better, albeit not much I guess. While it is true that there are more fusion breakpoints near the 5' end, the main...

Yes, it should be fine to use the regular STAR for fusion calling. Just make sure NOT to pass the FastQ containing the UMIs+barcodes to STAR. Only pass the transcript...

What I mean by that is that the alignments in the BAM file need to be marked as duplicates (flag `BAM_FDUP` set) in accordance with them having the same mapping...

> However STAR returned 0 bam files for many of my samples Probably, the reads need trimming of 10X adapters before alignment. Otherwise, STAR may discard them, because only part...

Is your plan to append the sequences of the fusion genes as separate contigs to the FastA file and then define GTF records for these new contigs.

In that case, it would be the easiest to run Arriba with `-I`. Arriba will then do it's best to report as much info about the fusion sequence as possible...

> what did you mean by: "define the entire fusion transcript as one big exon" in the .gtf file? Let's say one of the fusion sequences is 1000 bases long....