At HIMSS25, Shweta Maniar, Director of Strategic Industries for Healthcare & Life Sciences at Google, and Dr. Santosh Dixit, Chief Domain Expert at HCLS Innovation Labs at Persistent, spoke to Rebecca M. Knight from theCUBE, about harnessing actionable insights from a deluge of clinical data using knowledge graphs within the life sciences industry.
Monetizing clinical data requires correlating vast amounts of unstructured and structured multi-modal data to extract underlying relationships. For richer insights and relationships, clinicians and scientists must consider a diverse dataset that spans microscopy, pathology, genomics, family history, and other relevant areas. Done manually, this can prove overwhelming and lead to delays that in some circumstances could be life-threatening.
Persistent partnered with Google to create Pi-OmniKG, a knowledge graph-based solution that consolidates disparate, siloed, multi-modal clinical data to extract contextual relationships that may not be discernible instantly to the naked eye. The solution can incorporate as many data sources as needed, creating rich insights with explainable interrelationships that help establish trust with users, who may or may not have the technical expertise to run AI-led systems.
Built on top of the Google Cloud stack, Pi-OmniKG leverages BigQuery as the data warehouse to store premium upstream data, VertexAI as the AI engine that connects the dots to create a knowledge graph, and the Gemini toolkit to query the model. The Google stack enables enterprises to scale the solution according to their data needs, treating every data point as sacred, meaningful, and usable.
For instance, in precision medicine with hyper-targeted treatments, clinicians need to identify the direct relationships between a patient’s genomic profile and drug interactions. Pi-OmniKG can help create those connections by leveraging a vast amount of underlying data and generating actionable, timely insights that enable the creation of hyper-focused treatment plans with high degrees of explainability.
To scale the solution further, Persistent will collaborate with Google to enhance Pi-OmniKG’s core processes by incorporating newer and more diverse data sources, enabling users to create the underlying schema on the fly with just a prompt, and interacting with the knowledge graph more meaningfully through visual output.
Speakers–
Shweta Maniar, Director, Strategic Industries, Healthcare & Life Sciences at Google
Dr. Santosh Dixit, Chief Domain Expert at HCLS Innovation Labs at Persistent
Rebecca M. Knight, theCUBE
Learn more about our Google partnership.
Learn more about Pi-OmniKG.
