Skip to content
This repository has been archived by the owner on Feb 22, 2022. It is now read-only.

Latest commit

 

History

History
97 lines (65 loc) · 4.93 KB

README.md

File metadata and controls

97 lines (65 loc) · 4.93 KB
Raw

q2-dada2-CCS

Implement qiime dada denoise_ccs in q2-dada2 plugin for denoising Pacbio CCS long reads.

Test on qiime2-2021.4

Qiime2 v2022.2 has offically supported Pacbio CCS denoise.

Usage

  1. Overwrite the file __init__.py, plugin_setup.py, and _denoise.py in /path/to/your/conda/envs/qiime2-version/lib/python3.8/site-packages/q2_dada2/ with the file in this repo.
  2. put a new file run_dada_ccs.R in /path/to/your/conda/envs/qiime2-version/bin and use command chmod u+x ./run_dada_ccs.R to make sure you have execution authority.
  3. use command qiime dev refresh-cache to refresh your QIIME2 environment.

Detail

Changing following files in /path/to/your/conda/envs/qiime2-version/lib/python3.8/site-packages/q2_dada2/

  1. __init__.py : import a new function denoise_ccs from _denoise.py

  2. plugin_setup.py :

    1. Register the new function denoise_ccs based on denoise_pyro

    2. Add following explicit parameters :

      • front and adapter for primer removing
      • max_mismatch and indels for primer matching
      • min_len to filter the min length of CCS reads

    3. Change the default value of following explicits parameters :

      • trunc_len : default 0
      • min_fold_parent_over_abundance from 1 to 3.5
      • n_reads_learn from 250000 to 1000000 (same with denoise_single)

  3. _denoise.py :

    1. define the new function denoise_ccs based on denoise_pyro and _denoise_single :

      1. add new parameters :

        • front and adapter : string
        • max_mismatch : interger, default 2
        • indels : boolean, default False
        • min_len : interger, default 20

      2. change the default value of following parameters :

        • trunc_len : default 0
        • min_fold_parent_over_abundance from 1 to 3.5
        • n_reads_learn from 250000 to 1000000 (same with denoise_single)

      3. remove implicit parameters homopolymer_gap_penalty and band_size , re-define their default values in run_dada_ccs.R

      4. change the command line run_dada_ccs.R to work with previous changes :

        • add new temporary directory nop_fp for PRIMER REMOVING step
        • change the path of temporary directory filt_fp
        • add new arguments in run_dada_ccs.R for the new parameters in denoise_ccs

      5. return the stats log of primer removing step to function _denoise_helper

    2. change the function _denoise_helper for exposing the PRIMER REMOVING step stat in final denoise_stats.qza, including primer-removed and percentage of input primer-removed

    3. add new parameters to the _valid_inputs for paramters validation:

      • max_mismatch : _WHOLE_NUM
      • min_len : _WHOLE_NUM
      • front, adapter and indels : _SKIP

Then add a new Rscript run_dada_ccs.R based on run_dada_single.R in /path/to/your/conda/envs/qiime2-version/bin

  1. assign the new arguments and reorder all the arguments :

    • primer.removed.dir for the path of temporary directory in PRIMER REMOVING step , corresponding to nop_fp in denoise_ccs
    • primerF and primerR for primer sequences, corresponding to front and adapter in denoise_ccs
    • maxMismatch and indels for primer matching, corresponding to max_mismatch and indels in denoise_ccs
    • *minLen for min length CCS reads filtering, corresponding to min_len in denoise_ccs

  2. add a new PRIMER REMOVING step before TRIM AND FILTER step :

  3. modify the TRIM AND FILTER step :

    1. change the default value of implicit parameters rm.phix from TRUE to FALSE in function filterAndTrim
    2. add paramters in function filterAndTrim :
      • add minLen = minLen
      • minQ = 3

  4. modify the LEARN EROOR RATES step :

    1. change the default value of following implicit parameters :
      • errorEstimationFunction from dada2:::loessErrfun to dada2:::PacBioErrfun
      • add an explicit default value of BAND_SIZE to 32
      • remove the HOMOPOLYMER_GAP_PENALTY parameters so its default value change to NULL

  5. modify the REPORT READ FRACTIONS THROUGH PIPELINE step to add statistics of PRIMER REMOVING step in denoise_stats.qza

Enter qiime dev refresh-cache after changes those files.

Note: Still lack of auto test of those new parameters so make sure your input is valid.

All parameters used in those files are referenced from Callahan et al., 2019 and https://github.com/benjjneb/LRASManuscript

PR to qiime/q2-dada2, waiting for merging.