Skip to content

Latest commit

 

History

History
127 lines (90 loc) · 5.43 KB

README.md

File metadata and controls

127 lines (90 loc) · 5.43 KB
Raw

met_annot_enhancer

A set of script to proceed to metabolite annotation results enhancement using taxonomically and structurally informed processes.

Installation

1. Clone this repo

git clone https://github.com/mandelbrot-project/met_annot_enhancer.git

2. Set the correct environment

The requested package can be installed by creating a conda environment and activating it.

Use the following line for environment creation

conda env create -f environment.yml

And this one to activate it

conda activate met_annot_enhancer_env

Requirements

A FBMN job on GNPS

For now the scripts work by taking output of Feature Based Molecular Networking jobs. You can read more about FBMN here (https://ccms-ucsd.github.io/GNPSDocumentation/featurebasedmolecularnetworking/)

A metadata file containing taxonomical information

A specific requirement to proceed to taxonomical reweighting of your annotation results is that a metadata file with the taxonomical information of your samples is provided. See below for details.

2 columns of metadata are required for taxonomical reweighting:

  • a filename column (with 1 row / MS-file). Example: 2013901_AG_07098.mzML. ! The header for this column has to be filename.
  • an organism column (organism the sample was collected from). Example: Ambystoma mexicanum

This file is ideally uploaded when you relize your FBMN job. See here for details on the metadata file format (https://ccms-ucsd.github.io/GNPSDocumentation/metadata/)

If you didn't add it you can do this afterward. For this, the metadata file, in the form of a .tsv file, should be placed in data_in/yourjobid/metadata_table/metadata_file. The name of the file doesn't matter but it should be the only file in the folder.

Usage

1. Edit the .yaml file containing the job parameters

For this you can first copy the default.yaml file from here to there. Don't rename the file.

Now you can safely edit the file in configs/user_defined/default.yaml according to your needs. See below for a brief descrpition of each parameters

Parameters description:

options:

  • download_gnps_job: True
    • set to False it you already downloaded a GNPS FBMN
  • do_spectral_match: True
    • will perform MS2 matching using db_file_path .mgf
  • keep_lowest_taxon: False
    • for clarity un outputs, just keep the lowest taxon matched
  • output_plots: True
    • False if keep_lowest_taxon = True (to change)

paths:

  • gnps_job_id: 250536f4cb3e4f159e5ef67a3d024fac
    • The GNPS job id you want to annotate
  • project_name: your name
    • The name you want to give to your project, output resulst in data_out/project_name
  • metadata_path: db_metadata/210523_lotus_dnp_metadata.csv
    • Path to your spectral library file
  • db_file_path: db_spectra/LOTUS_DNP_ISDB.mgf
    • Path to the metadata of the spectral file
  • adducts_pos_path: data_loc/db_prepared_pos.tsv.gz
    • Path to the adducts file in pos mode
  • adducts_neg_path: data_loc/db_prepared_neg.tsv.gz
    • Path to the adducts file in neg mode

spectral_match_params:

  • parent_mz_tol: 0.01
    • the parent mass tolerance to use for spectral matching (in Da)
  • msms_mz_tol: 0.01
    • the msms mass tolerance to use for spectral matching (in Da)
  • min_cos: 0.3
    • the minimal cosine to use for spectral matching
  • min_peaks: 8
    • the minimal matching peaks number to use for spectral matching

repond_params:

  • Top_N_Sample: 0
    • Max number of contributors to take into account for taxo reponderation, set to 0 for all biosources where the feature is detected
  • top_to_output: 1
    • Top X for final ouput
  • ppm_tol: 2
    • ppm tolerance to be used for ms1 match
  • polarity: 'pos'
    • ion mode you are working with (pos or neg)
  • organism_header: 'ATTRIBUTE_Species'
    • Mandatory: header of your samples' organism in metadata file
  • var_one_header: 'SAMPLE_info'
    • Optional parameter
  • use_post_taxo: True
    • Set True if you want to use rank after taxonomical reweighting for consensus chemical class determination, else set to False
  • top_N_chemical_consistency: 15
    • if use_post_taxo = True: Top N to use for chemical consistency (annotation not in top N will be discared for component consensus determination)
  • file_extension: '.mzML'
    • MS filename extension (or any common pattern in all your MS filenames)
  • msfile_suffix: ' Peak area'
    • Suffix to remove in you quant.csv table to match your MS filenames in metadata table
  • min_score_ms1: 5
    • Minimum taxonomical score (7 = species, 6 = genus, 5 = family, ...)

2. Launch the job

From the home folder of this repository. In the activated conda environment.

python src/dev/nb.py

References

Description and original implementation of the taxonomically informed metabolite annotation process is available here https://doi.org/10.3389/FPLS.2019.01329, associated data have been deposited in the following OSF repository: https://osf.io/bvs6x/. A snapshot of the code at the time of publication is also available at https://github.com/oolonek/taxo_scorer.

An R based implementation of the metabolite annotation enhancing process is available here https://github.com/taxonomicallyinformedannotation/tima-r