Uniprot meta tool

this tool is designed to be useful and simple interface for UniProt.org knowlege base witch can be used to simply get protein metadata by its ID

NOTE: this package is in early stage of development so be careful!

installation

as simple as possible:

pip install uniprot-meta-tool

only dependencies are requests and jmespath

Examples

from uniprot_meta_tool import UniprotData

>>>up_data = UniprotData('P05067')
>>>up_data.name
'Amyloid-beta precursor protein'
>>>up_data.genes[0]
'APP'
>>>up_data.organism.common_name
'Human'

some util functions are also available:

get_molecular_functions(uniprot_data) -> List[str]:
returns list of molecular functions from keywords. accepts uniprot_id as string or UniprotData object

get_biological_processes(uniprot_data) -> List[str]:
returns list of biological processes from keywords.

get_pathways(uniprot_data: UniprotData) -> List[str]:
returns list of pathways (from Reactome DB using UniProt cross-reference)    

utility to enrich dataframes (requires Pandas to be installed):

from uniprot_meta_tool.dataframe import enrich_df

df = enrich_df(
    df, # pandas DataFrame
    'protein_id', # name of column contains UniProt ID
    ['gene', 'protein_name', 'organism'] # list of parameters that should be added
)

Raw search

You can reach raw metadata from UniprotData(u_id).raw_data property.

You can use JMESPath query language to get something from raw data using UniprotData(u_id).search() method

Configuration

You can set up basic parameters of how uniprot_meta_tool works with remote API and local storage, but be careful with this

from uniprot_meta_tool.data_parser import set_config
set_config(
    meta_cache_dir='/your/local/folder', # default is %tmp%/uniprot_meta
    use_cache=True, # if you set it to False, module will not store retrieved info to your disk 
    #                 and will perform new HTTP request every time you create UniprotData object
    uniprot_rest_url='https://rest.uniprot.org/uniprotkb/{}' # have no idea why you could want to change it, but you can
)

MR and feature requests

Feel free to suggest features using GitHub issues and to implement something you need if you want to!