AI is smart. But shouldn’t AI designed to serve retail customers with “loyal sushi customer” special offers be engineered to run at one level, with AI tasked with serving not just mission-critcal but also life-critical scientific AI use cases be even smarter? If the answer is yes (which it almost certainly is), then we will need to engineer that science-tasked AI with an unfathomably large data access channel that is as diverse as it is controlled.
The news follows on a string of recently signed partnerships between TetraScience and Google Cloud, Databricks and Nvidia, each of which support the startup’s ambitions to transform siloed, proprietary and unstructured scientific data so it can be used by engineers and scientists for AI and data apps to improve each step in drug development and manufacturing.
The specific collaboration [with Snowflake] will allow the sharing of scientific data between the Tetra Scientific Data and AI Cloud and Snowflake’s AI Data Cloud so that life sciences organizations can generate previously unattainable insights with diverse and multimodal datasets. This is said to improve their pace of operations and the quality of decision-making across the biopharma value chain.
Why Is Science Data Raw?
The Tetra scientific data and AI cloud is built to transform raw, siloed scientific data into analytics-ready and AI-native datasets. Why is scientific data so often raw and siloed? Because this is information that might sit on only one single extremely specialized piece of scientific equipment in any given research facility (so it’s a single solitary disconnected silo) and that machine itself will most likely be engineered to perform a deeply complex scientific analysis task, meaning little or no design effort goes into making sure it also has a slick data intelligence function set built into its firmware or software layer (so that’s the raw part) when it is set to work.
Either that… or it’s a facility where raw siloed data is the norm due to legacy data team practices. Or it’s both.
By combining TetraScience’s expertise in scientific data and workflows with the Snowflake AI Data Cloud, organizations are promised the ability to gain access to scientific data and the agility to develop advanced analytics and AI-driven use cases. Scientific teams can collaborate on data across groups and geographical boundaries, using Snowflakes’s security and privacy safeguards.
“Our collaboration with Snowflake puts unprecedented operational insights and AI-ready scientific data directly in the hands of laboratory and data scientists,” said Patrick Grady, CEO of TetraScience. “Whether developing custom dashboards or running advanced analytics or AI models, researchers can now focus on science and innovation rather than data preparation.
Assay Reports, All Sorts
Several pharmaceutical organizations using Snowflake say that they already realize the benefits of accessible, analytics-ready AI-native scientific data from TetraScience. This enables diverse use cases and digital initiatives across the biopharma trade, such as monitoring the health and utilization of scientific instruments to enhance lab efficiencies and automating the ingestion and labeling of “assay reports” for secure and streamlined collaboration… and (although the tech industry uses those terms too much) in biopharma, when lives depend upon it, you want your data to be secure and streamlined.
For completeness here, an assay report and the data that belongs to it is a foundational laboratory test that measures the amount or presence of a specific substance or other target entity. The same term is used in mining when prospectors are working to determine the presence of precious or other metallurgical substances (so many of us would know it from cowboy Western movies) and it is also used in medicine, environmental science, pharmacology and forensics
“The combination of TetraScience’s deep scientific expertise with Snowflake’s AI Data Cloud creates a new opportunity for customers to gain efficiencies and scalability,” said Lisa Arbogast, industry principal for life sciences at Snowflake.
The collaboration between these two companies encompasses joint technology development, co-marketing and co-selling initiatives.
Data, The Lifeblood
We hear a lot of talk from the IT industry about data being the “lifeblood of business” because organizations want to convince us that they are data-centric cloud-native operations. Paradoxically, this is rather more a case of data being the “business of lifeblood” as specialist firms like TetraScience work to accelerate scientific AI by designing and industrializing AI-native scientific datasets.
We might reasonably argue that this work is essential if we’re going to exert the same levels of AI and data analytics that we see every day in manufacturing, retail or commerce also being applied by the biopharma community. Wouldn’t it be more prudent to automatically convert raw scientific data into an open vendor-agnostic format so that we can maximize data’s value through best-of-breed analytics and AI applications? Well, yes, that’s exactly what TetraScience does with its AI-native scientific datasets designed to allow data science engineers in life sciences roles to consume harmonized data with scientific taxonomies and ontologies.
Next time you hear some keynote speaker tell you that data is the lifeblood of our modern existence, at least you can sit quietly back in your chair and think about how true those words really are.
