Duplex call for rerio model fails with "Could not find simplex model information"; simplex works fine

[muntean-alex]

Dear all,

I'm trying to run a duplex basecall with the use of the model res_dna_r10.4.1_e8.2_400bps_sup@2023-09-22_bacterial-methylation.

I try to run
dorado duplex ./rerio/dorado_models/res_dna_r10.4.1_e8.2_400bps_sup@2023-09-22_bacterial-methylation /mnt/storage/data/2024.10.11-pod5 > /mnt/storage/data/2024.10.11-rerio-duplex-sup.fastq --emit-fastq
and get

[2025-03-01 12:19:13.182] [info] Running: "duplex" "./rerio/dorado_models/res_dna_r10.4.1_e8.2_400bps_sup@2023-09-22_bacterial-methylation" "/mnt/storage/data/2024.10.11-pod5" "--emit-fastq" "--models-directory" "./rerio/basecall_models/"
[2025-03-01 12:19:13.212] [info] > No duplex pairs file provided, pairing will be performed automatically
[2025-03-01 12:19:13.405] [info]  - Note: FASTQ output is not recommended as not all data can be preserved.
[2025-03-01 12:19:13.408] [error] finalise() not called on a HtsFile.
[2025-03-01 12:19:13.408] [error] Could not find simplex model information from: res_dna_r10.4.1_e8.2_400bps_sup@2023-09-22_bacterial-methylation

While running a simple dorado basecaller command works.
I have tried downloading all dorado models to the same directory, but nothing seems to work.

Can you offer any advice?

Many thanks,
Alex

[klawrence26]

Hi Alex,

The Rerio bacterial-methylation basecalling model does not support duplex basecalling, as it is a research model rather than a major release model. For bacterial samples, we recommend using the v4.3 or v5.0 SUP release models in Dorado, which should give similar or better accuracies on bacterial genomes compared to the older Rerio model, and which both support duplex basecalling.

Best,
Katherine