How much reads per cell for fusion detection with Nanopore ?

Question

[Evan]

Hello,

I'm willing to detect fusion within cancer cells at Single-Cell level using Nanopore Sequencing. I wondered how many reads per cell would be preferable to detecting known fusion. Do some of you have experience or good resources to advise?

Thank you for your help!

Evan

[dthorbur]

The authors of the tool JAFFAL have estimated sensitivity of identifying fusion genes across of varying transcriptome coverages in RNAseq data. They found 98% of simulated fusions were detected in simulated data with 10X coverage of the transcriptome. But other factors like the size of the transcriptome will significantly affect the number of reads you'll need if you use the RNAseq approach.

I am unaware of methods and limitations to identify fusion genes using genomic DNA datasets.

[Evan]

Thank you for your reply! Looking forward to seeing the results once the sequencing will be done, I may update this post :)

[dthorbur]

Having looked a little more in depth into the requirements of JAFFAL, it's actually not recommended to be used with Nanopore due to the relatively high error rates in comparison with PacBio. I was hoping to use it on a Nanopore cDNA library, but now need to find alternatives.

I'll try to remember to update the post if I find something. Otherwise I may try to tune tools designed for Illumina reads instead. STAR-SEQR (fusion) and SLIDR-SLOPPR (SL-trans-splicing) being the two that I have used.

[Evan]

Thank you very much for the info. I will update this post too to keep the community updated. I saw there is this other tool fusionSeeker recently published. In our case, even if there is a high error rate, we have a know fusion present in every sample, so I expect it will still work, if not I will be very surprised but we never know.