Skip to content

Latest commit

 

History

History
40 lines (36 loc) · 3.02 KB

README.md

File metadata and controls

40 lines (36 loc) · 3.02 KB
Raw

Model Training

Preprocessing

Alignment and Indexing
  • Make sure minimap and samtools are in your PATH (source load_environment.sh from your home directory)
  • In order for signalAlign to find the corresponding fast5s and fastqs you have to split and index the fast5 files. You also need to align fastqs to reference and filter alignment files. Note: this will only work if you untar our fast5 files with their corresponding directory structure inside the fast5 directory and the fastq directory maintains the same directory structure. See end_to_end
    • bash /home/ubuntu/yeast_rrna_modification_detection/training/prep_ivt_and_native_data.sh /home/ubuntu/fast5 /home/ubuntu/fastq /home/ubuntu/yeast_rrna_modification_detection/training/reference/yeast_25S_18S.fa

IVT vs Wild Type Training

  • In order to run training you need to alter the paths in the config file train_ivt_native.config.json
    • Under samples there is one named "wild_type". You need to change the paths to the corresponding file within your path system 
      "positions_file": "/home/ubuntu/yeast_rrna_modification_detection/training/small_5mer/yeast_18S_25S_modified.positions",
"fast5_dirs": ["/home/ubuntu/"],
"bwa_reference": "/home/ubuntu/yeast_rrna_modification_detection/data_preprocessing/reference/yeast_25S_18S.fa",
"readdb": "/home/ubuntu/fastq/20210410_R941_WT_YPD/20210410_1830_MN20528_AGN282_f7dbe316/20210410_1830_MN20528_AGN282_f7dbe316.training.readdb",
"alignment_file": "/home/ubuntu/fastq/20210410_R941_WT_YPD/20210410_1830_MN20528_AGN282_f7dbe316/20210410_1830_MN20528_AGN282_f7dbe316.2308.sorted.bam",
    • Under samples there is one named "ivt". You need to change the paths to the corresponding file within your path system 
      "positions_file": "/home/ubuntu/yeast_rrna_modification_detection/training/small_5mer/yeast_18S_25S_canonical.positions",
"fast5_dirs": ["/home/ubuntu/"],
"bwa_reference": "/home/ubuntu/yeast_rrna_modification_detection/data_preprocessing/reference/yeast_25S_18S.fa",
"readdb": "/home/ubuntu/fastq/10142019_R941_IVT/20191014_1845_MN20528_FAL23261_9a8cb226/20191014_1845_MN20528_FAL23261_9a8cb226.training.readdb",
"alignment_file": "/home/ubuntu/fastq/10142019_R941_IVT/20191014_1845_MN20528_FAL23261_9a8cb226/20191014_1845_MN20528_FAL23261_9a8cb226.2308.sorted.bam",
    • Finally, you need to put the path to signalAlign's bin, the base model, output dir, job count (n threads), and the ambig model. 
      "path_to_bin": "/home/ubuntu/src/signalAlign/bin",
"template_hmm_model": "/home/ubuntu/yeast_rrna_modification_detection/training/small_5mer/rna_r94_5mer_ACGTabc.model",
"output_dir": "/home/ubuntu/training/ivt_vs_wt",
"ambig_model": "/home/ubuntu/yeast_rrna_modification_detection/training/small_5mer/small_variants.model",
"job_count": 8
  • Once your config file is set up, run the following command.
    • trainModels.py run --config /home/ubuntu/yeast_rrna_modification_detection/training/train_ivt_native.config.json