Precision medicine and the discovery of biomarkers

Precision medicine and the discovery of biomarkers

precision_medicine_biomarkers

In medicine, a Biomarker is a biological indicator, which can mark a state of a biology and is measurable. A Biomarker or a combination of few biomarkers can establish the biological state of an individual helping us understand the exact disease state and potential treatments. As a simple example, biomarker for fever is high temperature.

 

Biomarker discovery is important because there are many diseases in which biomarkers are still not identified. A perfect example of this is the ongoing pandemic, where COVID-19 is impacting different age groups differently. Not only that, reports of different kinds of symptoms from different geographies have also surfaced. For example, metabolomic biomarkers play an important role, because they can be useful in defining personalized therapies based on various factors such as demographics etc., i.e. precision medicine.

 

Biomarker discovery is essential for precision medicine. It enables healthcare providers to identify groups of individuals with similar diagnosis of the disease, prognosis of how the disease would proceed and its severity. They also allow the response of drugs to be observed in different sets of patients. The right biomarkers can help identify the right combination/s of drugs and therapies for patients with the same disease but different outcomes.

 

For instance, chemotherapy only works in a certain percentage of cancer patients. Many cancer patients undergo chemotherapy, spending large sums of money, but may see little to no improvement. Through biomarker and precision medicine, these patients can be segregated at an earlier stage and can be offered alternate therapies for a better chance of survival.

 

Limited knowledge about the right biomarkers holds back the true potential of recovery in a patient. Omics data sets such as genomics, proteomics, metabolomics hold a reservoir of information that can help in biomarker discovery. However, availability of data and finding the right set of data is a major challenge in life sciences.

 

Secondly, life sciences data is not only siloed but that which is publicly available is heterogeneous. For example, scientists may use different assumptions and parameters in defining the experiment protocol, making the results of their experiments incomparable. These disparate data sets often limit the scope of identifying a single factor as the biomarker.

 

Moreover, human biology is quite complex as well as dynamic. Studies collect data from hundreds of patients, however the factors brought into consideration for these studies are quite narrow in comparison to hundreds of different potential markers which can be found in the human body. Due to similar reasons, traditional biomarker discovery methods are often limited and isolated, biased with pre-defined hypotheses, and suffer from high false-positives.

 

Our approach

 

Innoplexus, in its aim to assist precision medicine, has developed a biomarker discovery method. We leverage artificial intelligence technologies to automate the method of biomarker discovery.

 

Innoplexus contextualizes data from thousands of publicly available life sciences sources, including clinical research, experimental studies, published chemicals and their preclinical studies, and siloed datasets. We use this data to generate an initial set of potential biomarkers and patient responsiveness.

 

biomarker-tool

 

A hypothesis is generated by examining data from patients, identifying genes with altered expression, analyzing survival outcomes, and connecting these outcomes to specific biomarkers. These biomarkers are then validated in systems biology by assessing their implications via the Innoplexus life sciences data ocean and our life sciences ontology.

 

Some of the advantages of our method are:

 

  • Integrative biomarker analysis from public, enterprise, and third-party data
  • Ontology-enabled research for scientifically-validated basis for unbiased hypotheses
  • Identification of biologically-optimal biomarkers, reducing false-positives
  • High biomarker prediction accuracy with extensive network-based biological validation

Author

Vatsal Agarwal

Vatsal Agarwal is the Vice President, Artificial Intelligence & Computational Linguistics at Innoplexus, with an additional role as Head of Partnerships. Prior to this, he had worked as the Vice President, Technology and Innovation and Subject Matter Expert of Artificial Intelligence, filing over 40 patent application and presentation of work in various journals and conferences. He works with development teams on planning implementation of AI technologies, ontologies, and NLP. After graduating from IIT Roorkee, Vatsal held positions at NextGen Invent Corporation, John Hopkins University, Tata Consultancy Services. He specializes in developing intellectual property, building teams, planning & executing processes, and is skilled in Artificial intelligence, Machine learning, Natural language processing, Big data, and Bioinformatics.

Download Technology Capabilities
close slider

Download Technology Capabilities