OxfordNuffieldWRH / ebi-tools-api

Unofficial package to query the EBI Tools REST API

ebi-tools-api

Utility module to asynchronously query the EBI Tools REST API

Only BLASTp and Needle endpoints are implemented, but the codebase may be useful to add other endpoints easily.

Install

pip install git+https://github.com/OxfordNuffieldWRH/ebi-tools-api

Usage

from ebi_tools_api import EBITools

ebi = EBITools(email='[e-mail]')

mouse_taxonomy_id = 10090
human_il6 = 'MNSFSTSAFGPVAFSLGLLLVLPAAFPAPVPPGEDSKDVAAPHRQPLTSSERIDKQIRYILDGISALRKETCNKSNMCESSKEALAENNLNLPKMAEKDGCFQSGFNEETCLVKIITGLLEFEVYLEYLQNRFESSEEQARAVQMSTKVLIQFLQKKAKNLDAITTPDPTTNASLLTKLQAQNQWLQDMTTHLILRSFKEFLQSSLRALRQM'

result = await ebi.blastp(
    sequence=human_il6,
    title='Search for human IL6 homologues in mouse',
    taxids=mouse_taxonomy_id
)
# return a DataFrame
result.hits

Simplified summary:

result.hits_simple

Underlying JSON data converted to Python dict:

result.data

Nice graphical summary for Jupyter notebooks:

result.summarize()

Caching

By default the results will be cached on disk in .ebi_tools_cache to minimise traffic to EBI tools. To clear cache, either:

• pass reset_cache=True to a query function (e.g. blastp(sequence=human_il6, reset_cache=True) preserving all other arguments as in original call
• remove the contents of this hidden directory to clear entire cache at once

Pass verbose=True to see cache status.

You can change the path to cache in EBITools constructor:

ebi = EBITools(email=my_email, cache_dir='some/path/to/cache/dir')

Timeout

After sending an initial query, the tool will check for the result every i seconds, where i increases by one after each failed attempt, up to a configurable maximum interval (default 5 seconds). By default 100 attempts will be made. This can be customised in the constructor:

ebi = EBITools(email=my_email, attempts_threshold=20, backoff_limit=10)