BLOG
Target identification is the first step along the drug discovery timeline. It might sound relatively simple in theory: identifying a biological target that plays a role in disease. However, in practice, target identification can be a complex process requiring tedious work and time. Knowledge graphs have emerged as powerful tools for target identification within the life sciences field. In this article, we’ll cover how knowledge graphs can be used to improve target identification to eventually find better targets, faster.
Many targets are initially identified via reviewing scientific literature and experiential data and searching through public databases and an organization’s own internal data. In all, researchers must navigate through an overwhelming amount of evidence. Extracting what’s relevant from this extensive body of knowledge requires significant effort and time and can be prone to human error or oversight.
Making up part of this data is the massive amounts of heterogeneous data from multiple sources, such as genomics, transcriptomics, proteomics, and KO libraries that life sciences organizations generate and store. Integrating and analyzing this vast amount of data manually, in combination with literature and public data, is a time-consuming and challenging task. However, integrating these data sources is essential in order to gain a comprehensive and holistic understanding of the complex biological systems involved and potential mode of action.
Additionally, data integration helps to overcome the limitations of individual datasets. Each data source provides a partial view of the complex biological landscape, and by combining multiple datasets, researchers can overcome biases, fill in knowledge gaps, and validate findings across different sources, enhancing the reliability and confidence in any identified targets.
Life sciences data is often stored in diverse formats, ranging from structured databases to unstructured text and images. Integrating these different formats and schemas requires substantial effort to harmonize and standardize the data to ultimately make it compatible for analysis.
To add to this complexity, incomplete or inconsistent data further complicates the integration process. Different data sources may have missing values, incomplete annotations, or conflicting information, which need to be carefully addressed and resolved. Ensuring data quality and consistency becomes crucial for reliable target identification.
On top of all this, even when data is integrated, it can still be hard to navigate and even harder to derive insights from. This is where knowledge graphs come in. Not only do they assist in optimized data integration, but they also enable better data exploration and analysis once data is integrated and made available.
In the context of life sciences, a knowledge graph is a structured representation of information that captures relationships between various entities such as genes, proteins, diseases, drugs, and biological processes. It is based on graph theory, where nodes represent entities and edges represent the relationships between them. By organizing and integrating diverse data from multiple sources, knowledge graphs provide a holistic view of the interconnectedness of biological information.
Efficient data harmonization
A knowledge graph’s semantic representation enables the harmonization of data from different sources by mapping them to a common schema. By assigning consistent labels, properties, and relationships to the entities in the graph, knowledge graphs enable seamless integration of heterogeneous data, driving more efficiencies for researchers working in target identification.
Enriching data
Knowledge graphs can also enrich data by providing contextual information. Researchers can contribute additional information, such as annotations or metadata, to the knowledge graph that would’ve otherwise remained isolated. This enrichment enhances the quality and depth of the integrated data, providing a more comprehensive and contextualized view. The additional context aids in the interpretation of the data and supports more informed decision-making during target identification.
Improved data exploration and analysis
When it comes to exploring and analyzing data for target identification, knowledge graphs excel at linking and connecting data by capturing the relationships between entities. Through these relationships, data points from various sources are connected within the graph, allowing for efficient traversal and exploration. For target identification, this means that genes, proteins, diseases, pathways, and other relevant entities can be interconnected, enabling researchers to navigate the graph and explore the connections between different elements so they can pull targets that meet their precise criteria.
DISQOVER, ONTOFORCE’s flagship product, is founded on semantic technology and an ontology-based knowledge graph. As a knowledge discovery platform, DISQOVER seamlessly connects an organization’s internal, siloed data with licensed data and public data in one easy-to-use, customizable platform, enabling efficient data exploration and analysis.
With DISQOVER, data integration is made easy thanks to its data ingestion engine, which simplifies the import, transformation, and integration process for all data sources. The extraction, transformation, and loading of data happens in one place, managed via a graphical web frontend, requiring only basic and non-specialized IT skills.
DISQOVER also pre-ingests many of the top public data sources for the life sciences industry, along with licensed data. This means that once an organization’s internal data is ingested into DISQOVER, it is automatically linked to external data, enabling more efficient target identification research that doesn’t require manual review of literature or reports.
Ontologies, cross-references, contextual information, and name entity mapping are some of the various techniques that are utilized to further streamline and harmonize the data within DISQOVER.
Not only does DISQOVER improve data ingestion for target identification purposes, but thanks to the platform’s powerful data visualizations and dashboards, it also improves how researchers explore and analyze data, ultimately accelerating the target identification process. DISQOVER analyzes millions of data points from internal and external sources to generate relevant information, gathering fragmented sources into a holistic, easily searchable interface. This makes finding the right knowledge, evidence, and supporting data significantly more efficient. As a result, researchers spend less time searching, and organizations can identify new targets more quickly and with greater accuracy – giving them a crucial competitive edge.
One of our customers is pioneering network pharmacology company specialized on computational drug discovery with a focus on developing RNA interference (“RNAi”) therapeutics. They are using DISQOVER to efficiently find relevant scientific information and explore new relationships across a vast amount of connected data sources in order to drive their target identification.
When building their knowledge graph, this company brought together biological and chemistry informatics, scientific literature, OMICS based experiments, patents, and their own proprietary data to create the knowledge graph within DISQOVER.
For this company, using a knowledge graph has been an ideal solution for overcoming the challenges related to missing or incomplete data, as the knowledge graph facilitates the visualization and discovery of non-obvious relationships between diseases, genes, and processes. Additionally, predicting edges in the knowledge graph using manual or automated walks (via machine learning) also helps supplement their data and improve their research-related predictions and decisions.
“DISQOVER makes building a knowledge graph much easier. It allowed us to focus on the actual data itself, to curate it into a form with as little bias as possible.” – Head of Discovery Biology
Learn more about how knowledge graphs can accelerate drug discovery for biotechs by watching our webinar on the topic >>>
The accelerated target identification process provided by knowledge graphs has transformative implications for life sciences companies. It allows researchers to more efficiently sift through vast amounts of data, narrowing down potential targets and reducing the time and resources required for experimental validation.
The speed and efficiency offered by DISQOVER specifically thanks to optimized data integration and intuitive data visualizations enables companies to iterate and explore a broader range of target possibilities, which can ultimately lead to improved success rates in drug discovery.
