Analysing the extent of cell type information present in Wikidata: A case study on PanglaoDB

This manuscript (permalink) was automatically generated from jvfe/project_panglaodb@a206cb7 on September 22, 2020.

Authors

• João Vitor Ferreira Cavalcante
  0000-0001-7513-7376 · jvfe
  Bioinformatics Multidisciplinary Environment, Federal University of Rio Grande do Norte

• Tiago Lubiana
  0000-0003-2473-2313 · lubianat
  Computational Systems Biology Laboratory, University of São Paulo

Abstract

Wikidata, a freely editable knowledge graph database, presents a great opportunity for the integration of biomedical knowledge, it’s well thought linked data model can improve significantly the handling and distribution of scientific information. On the other hand, Wikidata is still lacking in various aspects, in particular to what pertains to cell type information. This study aims to analyse how cell type knowledge is currently modelled in Wikidata and how it differs from other type of biological information, using, as a reference point, metadata from the well known single cell RNA sequencing database, PanglaoDB.

Keywords: wikidata, knowledge graph, cell type, ontology.

Introduction

Wikidata

Wikidata is an open, freely editable, knowledge graph database within the semantic web that stores knowledge across a multitude of domains, such as arts, history, chemistry and biology, using an item-property-value linked data model (Figure 1). It is easy to use and edit, by both humans and machines, with a rich web user interface and wrapper packages available in common programming languages such as R and Python. All the data within Wikidata is linked and inherently public domain, thus, it presents a great opportunity to make scientific data more FAIR (Findable, accessible, interoperable and reusable), as well as provides the necessary tools to curate and develop ontologies.

Several advances towards biological data integration and biological data analysis in Wikidata have been made before, yielding positive results [1] [2] and showcasing it’s potential for bioinformatics-related analyses, such as drug repurposing and ID conversion [2]. Wikidata has been proposed as a unified base to gather and distribute biomedical knowledge, with more than 50 000 human gene items indexed and hundreds of biomedical-related properties [3]. However, as of August 2020, cell type information is still very scarse, with only 264 items being categorized as “instances of cell types (Q189118)” (https://w.wiki/b2w), of those, only nine have a “Cell Ontology ID”[4] (P7963) associated, and most have a varying amount of statements (Table 1). As an additional problem, there are also 23 items being categorized as “instances of cell (Q7868)” (https://w.wiki/b2x), illustrating the absence of any formal data model.

PanglaoDB

PanglaoDB [5] is a public database that contains data and metadata on hundreds of single-cell RNA sequencing experiments, providing extensive information on cell types, genes and tissues, as well as manually and community curated cell type markers (Tables 2 and 3). It also provides a rich web user interface for easy data acquisition, including database dumps for bulk downloads.

Objectives

In this study, we aim to answer questions regarding the integration of biological data from PanglaoDB in Wikidata, analysing items such as cell types, genes and tissues. Some of the questions we gathered so far are:

• How many cell types in PanglaoDB are also present in Wikidata? How many of those items are exact matches?

• Of those that are exact matches, how many statements do they have associated? Are these items well annotated?

• How does the coverage of biological items differ within Wikidata? Are the items for tissues and genes better annotated? How so?

  • Do items with alternative identifiers, such as genes, have their alternative identifiers indexed within Wikidata?

In the end, we’ll have gathered and analysed enough data to formulate a report on the integration of this knowledge. We also intend to migrate the missing data, enriching Wikidata with more biological information.

Methodology

Data acquisition

Data from Wikidata will be acquired using the Wikidata Query Service and associated wrapper packages in Python, such as WikidataIntegrator and wikidata2df.

Data from PanglaoDB will be acquired through their web interface and metadata database dump repository.

All data used will be handled with commonly used Python data science packages, such as Pandas[6], Seaborn[7] and Jupyter[8].

Reconciliation

The metadata from PanglaoDB on cell types, genes, tissues (including germ layers) and organs will be matched to Wikidata items using the reconciler Python package, which is itself a wrapper around the well known OpenRefine reconciliation service, as well as manual intersections of both data sources. Data from the reconciliation service will be considered a match if the service returns a value of “match” equals to “True”.

Item quality assessment

Wikidata items will be assessed for their quality by their number of statements, which can be acquired via both the MediaWiki API and Wikidata’s own query service. And also by the presence of external identifiers, such as Ensembl Gene[9] and Entrez Gene[10] IDs for genes, Cell Ontology[4] IDs for cell types and Uberon[11] IDs for organs and tissues.

References