dorado icon indicating copy to clipboard operation
dorado copied to clipboard

Failed to get modification model

Open KunFang93 opened this issue 1 year ago • 1 comments

Hi,

I met a error when run dorado basecaller hac,5mCG_5hmCG,6mA output_01.pod5 > test2.bam

(base) kfang@icrc4lcga001:~/data/kfang/NucPatternOri/ONTdata$ dorado basecaller hac,5mCG_5hmCG,6mA output_01.pod5 > test2.bam
[2024-09-27 21:23:11.650] [info] Running: "basecaller" "hac,5mCG_5hmCG,6mA" "output_01.pod5"
[2024-09-27 21:23:11.792] [info]  - downloading [email protected] with httplib
[2024-09-27 21:23:12.235] [error] Failed to get modification model
[2024-09-27 21:23:12.235] [info] Found 2 modification models without mods variant: 6mA
[2024-09-27 21:23:12.235] [info] - [email protected]_5mCG_5hmCG@v0 - mods variant: 5mCG_5hmCG
[2024-09-27 21:23:12.236] [info] - [email protected][email protected] - mods variant: 5mCG
[2024-09-27 21:23:12.236] [error] No matches for chemistry: dna_r9.4.1_e8, model_variant: hac, version: v3.3.0, mods_variant: 6mA
[2024-09-27 21:23:12.236] [error] finalise() not called on a HtsFile.

I wondered how I could solve this issue? Thanks in advance!

Best, Kun

KunFang93 avatar Sep 27 '24 21:09 KunFang93

Hi @KunFang93,

There is no 6mA mods model for the r9.4.1 condition which is what the information messages are trying to communicate.

Failed to get modification model
Found 2 modification models without mods variant: 6mA
- [email protected]_5mCG_5hmCG@v0 - mods variant: 5mCG_5hmCG
 - [email protected][email protected] - mods variant: 5mCG

Best regards, Rich

HalfPhoton avatar Sep 28 '24 15:09 HalfPhoton

Thanks! The code I tried is from example code show in GitHub page under Modified basecalling section, I am not sure why it cannot automatically refer to model works for 6mA. I will specify model for 6mA. I wondered if these codes is correct format:

 dorado basecaller [email protected] output_01.pod5 --modified-bases 6mA > test2.bam

Thanks again!

KunFang93 avatar Sep 30 '24 15:09 KunFang93

@KunFang93 ,

Those commands looks correct for r10.4.1 data. Note that your output above appears to indicate that your data is r9.4.1 - basecalling with the r10.4.1 model is not recommended as it will not produce accurate results for this data.

malton-ont avatar Oct 01 '24 12:10 malton-ont