NDEx NCI-PID content loader

Python application that loads NCI-PID data into NDEx

This tool downloads OWL files containing NCI-PID data from: ftp://ftp.ndexbio.org/NCI_PID_BIOPAX_2016-06-08-PC2v8-API/ and performs the following operations:

1) OWL files are converted to extended SIF format using Paxtools and the SIF file is loaded into a network

2) A node attribute named type is added to each node and is set to one of the following

by extracting its value from PARTICIPANT_TYPE column in SIF file:

• protein (originally ProteinReference)
• smallmolecule (originally SmallMoleculeReference)
• proteinfamily (set if node name has family and was a protein)
• RnaReference (original value)
• ProteinReference;SmallMoleculeReference (original value)

3) A node attribute named alias is added to each node and is loaded from UNIFICATION_XREF column in SIF file which is split by ; into a list. Each element of this list is prefixed with uniprot: and t first element is set as the represents value in node and removed from the alias attribute. If after removal, the alias attribute value is empty, it is removed.

4) In SIF file INTERACTION_TYPE defines edge interaction type and INTERACTION_PUBMED_ID define value of citation edge attribute. The values in citation edge attribute are prefixed with pubmed: Once loaded redundant edges are removed following these conventions:

• neighbor-of edges are removed
• controls-state-of edges are removed if another edge connecting same nodes has one of the following interactions: controls-state-change-of, controls-transport-of, controls-phosphorylation-of, controls-expression-of

NOTE: If above results in orphaned nodes, those nodes are removed as well

5) An edge attribute named directed is set to True if edge interaction type is one of the following (otherwise its set to False)

controls-state-change-of
controls-transport-of
controls-phosphorylation-of
controls-expression-of
catalysis-precedes
controls-production-of
controls-transport-of-chemical
chemical-affects
used-to-produce

6) If node name matches represents value in node (with uniprot: prefix added) then the node name is replaced with gene symbol from gene_symbol_mapping.json

7) If node name starts with CHEBI then node name is replaced with value of PARTICIPANT_NAME from SIF column

8) If node represents value starts with chebi:CHEBI the chebi: is removed

9) If _HUMAN in SIF file PARTICIPANT_NAME column for a given node then this value is replaced by doing a lookup in gene_symbol_mapping.json, unless value in lookup is - in which case original name is left

10) Any node with family node name is changed as follows if a lookup of node name against gene_symbol_mapping.json returns one or more genes

• Node attribute named member is added and set to list of genes found in lookup in gene_symbol_mapping.json
• Node attribute named type is changed to proteinfamily

11) Changed in 5.0.0. For each network all proteinfamily nodes are examined and if any members exist

as separate nodes, those nodes are removed and their edges are shifted to the corresponding proteinfamily node. Duplicate edges are removed and other edges are merged if interaction and directed values are the same. In the case of a merge citation field values are merged into a new unique list.

12) The following network attributes are set

• name set to name of OWL file with .owl.gz suffix removed except for PathwayCommons.8.NCI_PID.BIOPAX which is renamed to NCI PID - Complete Interactions
• author (from Curated By column in networkattributes.tsv)
• labels (from PID column in networkattributes.tsv)
• organism is pulled from organism attribute of style.cx
• prov:wasGeneratedBy is set to html link to this repo with text ndexncipidloader <VERSION> (example: ndexncipidloader 1.2.0)
• prov:wasDerivedFrom is set to full path to OWL file on ftp site
• reviewers (from Reviewed By column in networkattributes.tsv)
• version is set to Abbreviated month-year (example: MAY-2019)
• description is pulled from description attribute of style.cx except for NCI PID - Complete Interactions which has a hardcoded description set to This network includes all interactions of the individual NCI-PID pathways.
• networkType is set to list of string with single entry pathway except for NCI PID - Complete Interactions which also includes interactome
• __iconurl is set to value of --iconurl flag (currently defaulting to http://search.ndexbio.org/static/media/ndex-logo.04d7bf44.svg)
• __normalizationversion is set to 0.1

13) By default each network is made public with full indexed and showcased (visible in user's home network list page)

NOTE: gene_symbol_mapping.json was originally extracted from here but the gene families were updated by calling ndexloadncipid.py --getfamilies sifdir/ which calls https://mygene.info via biothings Python client

Dependencies

• ndex2
• ndexutil
• biothings_client
• requests
• pandas
• py4cytoscape

Compatibility

• Python 3.6+

Installation

git clone https://github.com/ndexcontent/ndexncipidloader
cd ndexncipidloader
make dist
pip install dist/ndexncipidloader*whl

Configuration

The ndexloadncipid.py requires a configuration file in the following format be created. The default path for this configuration is ~/.ndexutils.conf but can be overridden with --conf flag.

Format of configuration file

[<value in --profile (default ndexncipidloader)>]

user = <NDEx username>
password = <NDEx password>
server = <NDEx server(omit http) ie public.ndexbio.org>

Example configuration file

[ncipid_dev]

user = joe123
password = somepassword123
server = dev.ndexbio.org

Required external tool

Paxtools is needed to convert the OWL files to SIF format.

Please download paxtools.jar (http://www.biopax.org/Paxtools/) (requires Java 8+) and put in current working directory

Or specify path to paxtools.jar with --paxtools flag on loadnexncipidloader.py

Usage

For more information invoke ndexloadncipid.py -h

Example usage

This example assumes a valid configuration file with paxtools.jar in the working directory.

ndexloadncipid.py sif

Example usage with sif files already downloaded

This example assumes a valid configuration file and the SIF files are located in sif/ directory

ndexloadncipid.py --skipdownload sif

Credits

This package was created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.