BCOR:CCNB3 fusions not detected

[Dhwani-2410]

Hello @nadiadavidson,

We have an FFPE sample that was positive for BCOR-CCNB3 fusion. We have performed RNA sequencing for the same using nanopore technologies. I was not able to get any fusions using minimap2 approaches(JAFFAL.groovy) but I got the results using the bowtie2 aligner (JAFFA_direct.groovy). But since the data was generated using nanopore, bowtie2 was not able to handle the data. After 5-6 failures of running JAFFAL.groovy, I was able to get the results and I assume this is not a sustainable way to get fusion results.

Using minimap2 approaches (JAFFAL.groovy), the intermediate files have reads associated with BCOR fusion.

sample.fusions.fa

39d6c082-bcd9-4bcc-abb3-a13037b3f396 runid=f4c45bdde2fc4f7373bea945ed566b3b4b1fc656 read=41170 ch=467 start_time=2023-06-21T15:34:54.342519+05:30 flow_cell_id=FAX09215 protocol_group_id=324848_TRANSCRIPTOME sample_id=324848 barcode=barcode03 barcode_alias=barcode03 parent_read_id=39d6c082-bcd9-4bcc-abb3-a13037b3f396 basecall_model_version_id=2021-05-17_dna_r9.4.1_minion_384_d37a2ab9 ATTTGGTGTTTATGGATGTCGCCTACCGTGACGAGTCTTGTGTCCCAGTTACCAGGACTTGCCGCCGCTCTATCTTCAGAGGAGGGGCTGTAGCTTTTCCAGTAAGGACACCTCATCATGATTTGAGCACTTTTTTAAAGATAATGACTTATTGATAAGGGTTTCATCCTCAATAGTGGGTTCCTCCTTTAAAACTAATGGCTCTGAAGAGAGATGCCTCCTCAGTGTTAGGTGTTTTGGAGGTGGTGGATATGTCTAGAATAGTGGTTTCTCCATAATGTTTGTACCACGGTAGTAGAGGCTACTACTGGTGTGACTTCCAGCTTATGCCAGTAGTTGTCTGAGGCCAGATCACTGGGGTGGAGCCACTCTACAGAGAAACCCAAACGTTAAGGCTAAGTGGAATCTGAAGCTTTCTATCAGCACTTTCTACAACCAGGAAACCTGGTAACTGGGACACAAGACTCGTCAC

sample.txt

39d6c082-bcd9-4bcc-abb3-a13037b3f396 328 333 BCOR:CCNB3 472 hg38_wgEncodeGencodeCompV22_ENST00000342274.7__range=chrX:40051248-40177329__5'pad=0__3'pad=0__strand=-__repeatMasking=none 5526 hg38_wgEncodeGencodeCompV22_ENST00000276014.10__range=chrX:50284536-50351910__5'pad=0__3'pad=0__strand=+__repeatMasking=none 451 332 329 -

sample.paf

39d6c082-bcd9-4bcc-abb3-a13037b3f396 472 329 385 -hg38_wgEncodeGencodeCompV22_ENST00000342274.7__range=chrX:40051248-40177329__5'pad=0__3'pad=0__strand=-__repeatMasking=none 6390 5470 5526 54 56 0 NM:i:2 ms:i:100 AS:i:100 nn:i:0 tp:A:S cm:i:6 s1:i:41 de:f:0.0357 rl:i:0 cg:Z:56M

But no results in sample.3gene_summary.txt, sample.3gene_reads and JAFFAL_results.csv files

How do I troubleshoot this? since this positive control was validated by orthogonal methods for BCOR-CCNB3 fusions?

[nadiadavidson]

Sorry for the VERY slow reply to this. I assume probably not relevant anymore, but I will answer for the sake of others with similar issues.

JAFFAL (and JAFFA) perform a number of filtering steps which are designed to remove false positives, and like any approaches, they will sometimes remove true fusion reads. Your BCOR-CCNB3 fusion (thanks for sharing the sequence btw), shares a property that's common to many false positives - one of the breakpoints hits the middle of an exon rather than exon edge. This doesn't mean that it won't get reported, but it does mean that two or more fusion reads are needed, and it will be reported as Low Confidence. Due to the high error rate with ONT data, you can sometimes need even more than two reports for a fusion like this to be reported.

However, if you know in advance which fusions you expect, using the intermediate (unfiltered) files is perfectly okay and can increase detection sensitivity. These files just shouldn't be used for unbiased discovery of novel fusions.