Public-Health-Bioinformatics/FastaUploader: A tool for batch processing large fasta files and accompanying metadata table to upload to repositories via API

A tool for batch processing large fasta files and accompanying metadata table to repositories via API

The python script breaks large fasta files (e.g. 500mb) and related (one-to-one) tab-delimited sample contextual data into smaller batches of 1000 or some specified # of records which can then be uploaded to a given sequence repository if an API endpoint is selected. Currently there is one option for the API interface: VirusSeq.

This tool is developed by the SFU Centre for Infectious Disease Epidemiology and One Health in conjunction with VirusSeq and it works well with DataHarmonizer!

Authors: Damion Dooley, Nithu Sara John


Given a fasta file and a sample metadata file with a column that matches to fasta file record identifiers, break both into respective sets of smaller batches of records which are submitted to an API for processing.

Processing is three step:

  1. Construct batches of files. Since only two files are read and parsed in one go,
    processing of them is reliable after that point, so no further error reporting
    is needed during the batch file generation process.

    1. Importantly, if rerunning, this step will be skipped unless -f –force parameter is run. Currently input files are still required in this case.
  2. IF API option is included, submit each batch to API, wait for it to finish
    or error out (capture error report) and proceed to next batch.

    1. Some types of error trigger sudden death, i.e. sys.exit() because they would also occur in subsequent API batch calls. For example missing tabular data column names will trigger an exit. Once resolved, rerun with -f to force regeneration of output files.
  3. Report on processing status of existing API requests. Some may be queued, others may have been processed successfully, and others may have line-by-line errors in field content that must be addressed. Revise given batch files and run again.

Requires Biopython and Requests modules

  • “pip install biopython”
  • “pip install requests”


Run the command in a folder with the appropriate input files, and output files can be generated there too. Rerun it in the same folder to incrementally fix any submission errors and then restart submission.

python [options]


-h, --help
  show this help message and exit.
  provide a fasta file name.
  provide a COMMA delimited sample contextual data file name.
  provide a TAB delimited sample contextual data file name.
-b BATCH, --batch=BATCH
  provide number of fasta records to include in each batch. Default is 1000.
  provide an output file name/path.
  provide the metadata field name to match to fasta record identifier.
-a API, --api=API     
  provide the target API to send data too.  A batch submission job will be initiated for it. Default is "VirusSeq_Portal".
  an API user token is required for API access.
-r, --reset
  Regenerate all batch files and begin API resubmission process even if batch files already exist under given output file pattern.
-d, --dev
  Test against a development server rather than live one.  Provide an API endpoint URL.
 -s, --short
  Report up to given # of fasta record related errors for each batch submission.  Useful for taking care of repeated errors first based on first instance.

For example:

python -f “consensus_final.fasta” -c “final set 1.csv” -k “fasta header name” -u [enter API key here] -s 2

Read more here: Source link