Repository navigation
robin test on provided dataset ROBIN_test_set_A #120
Description
Activity
Hi,
Firstly - please try switching to the dev branch and running the test again.
It should not take three days - it should take a few moments - so something isn't correct.
The input bed file you are using is not correct - you should use the file robin/src/robin/resources/panel_11092024_5kb_pad.bed
This may be causing the slow down.
Can you see any errors in the log file you have written out? My suspicion is that the models have not been downloaded properly.
thks for the quick reply,
I switched to dev branch and changed the bed file thanks for that, i have this error at the begining2025-05-16 17:00:08,440 - robin.subpages.base_analysis - ERROR - Error processing BAM files: Exception in subprocess: Type: UnpicklingError Message: invalid load key, 'v'. Traceback (most recent call last): File "/home/aliouat/miniconda3/envs/robin/lib/python3.9/site-packages/nicegui/run.py", line 41, in safe_callback return callback(*args, **kwargs) File "/home/aliouat/miniconda3/envs/robin/lib/python3.9/site-packages/robin/subpages/NanoDX_object.py", line 170, in classification NN = NN_classifier(modelfile) File "/home/aliouat/miniconda3/envs/robin/lib/python3.9/site-packages/robin/submodules/nanoDX/workflow/scripts/NN_model.py", line 30, in __init__ model_files = pickle.load(f) _pickle.UnpicklingError: invalid load key, 'v'.in the directory robin/src/robin/models/
drwx------ 2 bioinfo bioinfo 102 avril 22 14:39 ./ drwx------ 12 bioinfo bioinfo 298 avril 22 14:39 ../ -rw------- 1 bioinfo bioinfo 138089639 mai 12 10:48 Capper_et_al_NN.pkl -rw------- 1 bioinfo bioinfo 1777245358 mai 12 10:49 general.zip -rw------- 1 bioinfo bioinfo 113 avril 22 14:39 __init__.py -rw------- 1 bioinfo bioinfo 203907645 mai 12 10:48 pancan_devel_v5i_NN.pklOk. It's going to take me a little while to check that as I am travelling. Will update you asap.
Reacted by aamyra
hi ,
we are testing robin pipeline on a Virtual machine (4 threads , 8 Go RAM), on the dataset provided in the issue #97,
https://github.com/LooseLab/ROBIN_test_set_A
we have robin version 0.1.7
The issue is that the pipeline is running for three days, but i still have no classification of CNS tumors neither with strugeon or nanoDX, how long sould it take for the pipeline to complete a classification ?
we also run sturgeon on the same datas, and we obtain a different classification, with halfless probes found in your dataet:
could you please have a look to the code and tell me if im missing somthing?
`
#samtools merge -@ 10 -o robintot.bam -c -b testmergebamrobin.tsv
#samtools index robintot.bam
#modkit adjust-mods --convert h m robintot.bam robintot_adjustmods.bam
#modkit extract full robintot_adjustmods.bam robintot_adjustmods_extractfull.txt
#sturgeon inputtobed -i robintot_adjustmods_extractfull.txt -o /testSturgeon/ -s modkit
#sturgeon predict -i merged_probes_methyl_calls.bed -o /testSturgeon/ --model-files /models_sturgeon/general.zip --plot-results
`
thank you for your help