What is Insitro?
Insitro is a Series B biotech company in the San Francisco Bay Area, founded just a couple of years ago in 2018, working on drug development using machine learning (ML) and advanced AI-powered predictive algorithms. Daphne Koller, the founder of Insitro, is a world leader in bioengineering and computational technologies with experience across academic as well as entrepreneurial endeavors.
Since the pharma industry experiences quite a bit of trial and error with the development of drugs and medicine, Insitro has adapted using better computational techniques like machine learning to both predict the probability of successful drug prescriptions and to avoid failures.
Insitro’s cell-based disease models, predictive insights, population-scale data approaches are critical to its success. Image Credit: Insitro.
Insitro’s technology helps to predict the appropriate path (among multiple options) for drug development. It integrates ML deeply in biology to train models on extensive datasets—from both the given patients and the biological aspects of the problem. Insitro is focused on utilising personnel with overlapping backgrounds in biology and machine learning to achieve advancements in cell biology and bioengineering. The data-driven approach does not rely on available organic datasets but also generates large and information-rich optimised datasets best suitable for ML algorithms.
Who is Daphne Koller?
Daphne Koller was a professor of Computer Science at Stanford University for 18 years and still continues her association with the school. She co-founded Coursera—one of the most famous MOOC-based learning platforms—in 2012 and served as the platform's co-CEO until 2014 and as the president until 2016. She also worked as the chief operating officer of Calico Labs, an Alphabet-owned Silicon Valley company that is working to develop an understanding of the ageing process. She featured on the list of TIME magazine’s 100 most influential people in 2012—one of the most notable among the numerous awards to her name.
She believes in working on blurring the boundaries between machine learning and biomedicine. It was in her quest to ensure that machine learning has more impactful results in biomedicine that she founded Insitro Inc. in 2018. This was in the interest of both designing novel techniques for medicine development and finding solutions to the drug-development R&D efforts. Insitro aims to address the lack of high-quality drug development-focused medical data and the lack of professionals who can focus on two major domains, instead of one: medicine and computer science.
Her exceptional skills, knowledge, and entrepreneurial record establish her as a very trusted and respected founder whose ideas inspire belief from the public. Her academic and professional accolades make her capable of running a company solving such issues.
Insitro’s Funding Round
Founded only two years ago in 2018, Insitro has raised appreciable amounts of money in venture funding to accelerate the company’s success and growth. After raising Series A funding in April 2019, the company just raised $143 million in an oversubscribed Series B round that closed on the 26th of May, 2020.
Silicon Valley-based company Andreessen Horowitz was the lead investor for the round along with several new and returning investors. The new investors included Canada Pension Plan Investment Board and funds managed by T. Rowe Price Associates Inc, Casdin Capital, WuXi AppTec’s Corporate Venture Fund, BlackRock, HOF Capital, and other unannounced investors. The returning investors include Two Sigma Ventures, GV, Third Rock Ventures, Foresite Capital, ARCH Venture Partners, and Alexandria Venture Investments.
Left to right: the logos of investment funds HOF Capital, Third Rock Ventures, Two Sigma Ventures, CPP Investment Board. Image Credit: Insitro.
Koller and Insitro acknowledged the need to put together a team suitable for bridging the gap between biology and computer science and thus onboarded a team of exceptional scientists, engineers, and drugmakers. The company’s strategy involves identifying the disease models then building predictive algorithms for different drugs under trial, before ultimately accelerating the development towards more favourable contenders.
Insitro has built a platform to develop—using different computational techniques—large disease-relevant datasets with enough representation for the human population. The company believes that AI algorithms (mainly those that are supervised learning-based) can only be meaningful to the field of medicine if the datasets are large enough to cover all variations while remaining unbiased. Applying genetic, phenotypic, and clinical data on machine learning methods help researchers to understand the foundational aspects of human biology—thus enhancing data-driven predictions and their approach.
A three-part process pipeline, which culminates in ‘Prediction Algorithms’.
The drug development process follows inducing human cells into pluripotent stem cells (iPSCs) capable of changing into any human body cell. Insitro’s understanding of disease architecture helps it to exploit induced iPSCs, genome editing, and high content cellular phenotyping techniques for disease model development. These models are highly accurate as they encode multi-dimensional human diversity thus producing very reliable results.
Using ML and the information-rich datasets, scientists can identify the differences between healthy and affected cells, the finest granularity, thereby developing models for the disease. Knowledge about disease models based on extensive data creates more stable predictions of that disease’s behaviour.
The predictive model is a combination of in-vitro cellular systems and in silico ML models capable of both identifying new diseases and finding the most potent drug-based solutions. The results enable better drug design processes and clinical development strategies. Further nuances of the approach include the identification of patients on the basis of their genetics, using several biomarkers and segmenting them.
Insitro Moving Forward
The company’s most significant partnership with Gilead instilled more belief in investors in Koller’s team and their ability to blur the boundaries between life sciences and data science. It is a three-year contract with a view to develop novel treatments for non-alcoholic steatohepatitis with an initial 15-million-dollar partnership upfront.
The industry strongly believes in the abilities of Daphne Koller and the company’s performance reinforces the same. A two-year-old company creating such remarkable stir is remarkable, and hopefully, the company lives up to the expectations of a safer and medically sound future.