Model for BINDINGKIT=103-496-900 not available. Unable to run predictions.

[tbenavi1]

Hello, I ran the ft predict-m6a command and received the following error:

[2025-02-18T20:08:48Z ERROR fibertools_rs::utils::bio_io] Model for BINDINGKIT=103-496-900 not available. Unable to run predictions.

I am using ft version v0.6.4. I am on Linux. Perhaps this is because we are using the new Revio SPRQ polymerase kit. Thank you.

[mrvollger]

We are recommending that users use the new PacBio m6A caller for this new chemisty. It has very similar performance to fibertools on older chemistries, and we don't have great training data to make a new model for fibertools with this new chemistry just yet:
https://bioconda.github.io/recipes/pbjasmine/README.html

Just be sure to use v2.4.0 or later.

[tbenavi1]

Thank you. It looks like with version 13.3 of the Revio software, 6mA are now called natively on the sequencing machine and are present in the MM and ML tags of the unaligned bam file. So, there is no need to run pbjasmine (and it looks like it might be deprecated based on the 404 error for https://github.com/PacificBiosciences/jasmine).

[mrvollger]

Leaving this open so others can see at least for a while.

I have also been told the pbjasmine is combing back "soon" but in the meantime the bioconda install is still there.

[tbenavi1]

Makes sense. For anyone reading, I tried running pbjasmine on the unaligned bam file, and it didn't work. So, I believe if you have the updated Revio software, you should skip the pbjasmine step. That's my best understanding at this moment, but I will update if I learn anything further.