Causaly: Empowering biomedical researchers with AI using Google Cloud

About Causaly

Causaly uses artificial intelligence to rapidly read, understand, and interpret vast databases of biomedical knowledge. Its platform surfaces evidence from 30 million biomedical publications in seconds, enabling researchers to rapidly map epidemiology data, biomarkers, genes, molecular targets, and identify potential treatment options.

Industries: Healthcare, Life Sciences
Location: United Kingdom

Tell us your challenge. We're here to help.

Contact us

Causaly developed a powerful research tool that enables biomedical scientists to gain rapid research insights using Google Cloud technology.

Google Cloud results

  • Enables rapid update cycles to iterate and operate AI-powered tools
  • Supports fast software development with low entry barriers
  • Offers early stage technical and development support to help startups scale their organization

Hosts application and production environments in one place

Scientific research is being produced at a faster rate than ever before. Between 2009 and 2019, global research output increased by about four percent year-on-year across science and engineering. While this tidal wave of research is leading to breakthroughs in everything from digital technologies to biomedicine, for researchers, the work of gathering sources, data, and evidence remains as painstaking as ever.

However, cloud-based machine learning is creating new possibilities for researchers to access the right sources quickly. Founded in 2017 by academics working in biomedicine, Causaly was launched to do just that. Causaly deploys artificial intelligence to rapidly read, understand, and interpret vast databases of biomedical knowledge from more than 30 million biomedical publications in seconds.

“One of the key factors in choosing our cloud provider was how much prior expertise we would need to build our first iteration of Causaly. Google Cloud had a lower barrier in that regard. My co-founders and I weren’t DevOps or cloud engineers, but we were able to start building pretty quickly.”

Artur Saudabayev, co-founder and CTO, Causaly

“As a researcher, I was relying on Google Scholar and IEEE, among other platforms, for years to gather evidence and information,” says Artur Saudabayev, Causaly’s co-founder and CTO. “With the development of natural language processing, my co-founders and I thought it must be possible to extract information from raw text and put it into a data structure that would allow people to ask complex questions, do their work faster, and quickly gain the insights they need to empower their decision-making.”

To realize its ambitions, Causaly built a powerful machine learning system capable of extracting relationships from unstructured text data semantically. Its algorithms are able to not only understand causal relationships between data points, but also convert them into evidential statements. Causaly then started looking for the right tools to build an application that would enable researchers from all over the globe to access the resulting insights. They selected a range of Google Cloud solutions for the job.

“One of the key factors in choosing our cloud provider was how much prior expertise we would need to build our first iteration of Causaly,” explains Saudabayev. “Google Cloud had a lower barrier for getting started quickly. My co-founders and I weren’t DevOps experts or cloud engineers, but we were able to start building pretty quickly.”

Bringing a startup vision to life with Google Cloud

To build and iterate its platform, Causaly started its journey with Compute Engine, a customizable compute service that enables developers to quickly create and run resource-intensive virtual machines. Causaly uses these machines to run its natural language processing pipeline and host the graph database containing the results of its analyses. Today, Causaly uses Compute Engine exclusively for its database stack as well as numerous internal services.

Meanwhile, in its early stages of development, Causaly entered the Google for Startups program, which provided technical support and compute credits to Causaly as it built its platform.

“Google was very responsive and helpful in terms of working with us when we were an early stage startup, with only two or three team members,” says Saudabayev. “The Google team supported us not only as technology partners, but also as strategic advisors who played a role in helping us grow as an organization.”

From there, Causaly used App Engine to bring together 240 million evidential statements and build its powerful knowledge graph that enables researchers to rapidly map epidemiology data, biomarkers, genes, and molecular targets and identify potential treatment options.

“What we love about Google Cloud is that its ecosystem addresses many heterogeneous business requirements, from scaling capacity with Compute Engine to security with Google Cloud Armor, and much more. It’s the fact that we can build and iterate tools that can communicate with each other in one place.”

Artur Saudabayev, co-founder and CTO, Causaly

Bringing more knowledge to researchers, quickly

One of the biggest challenges Causaly faced when it came to mapping biomedical research is navigating the sheer volume of data available across more than 30 million academic papers, with this number growing by around 100,000 documents every single month.

“We believe that today's approach to working with literature and evidence doesn't inherently scale with the pace of data growth,” says Saudabayev. “We want to be a next-generation tool that can manage this growth for users. To do that, our technology needs to grow and accommodate large volumes of data that are decentralized and siloed in different domains and databases.”

To manage its data, Causaly utilises a hybrid multicloud environment which includes Cloud Storage to enable collaboration with its partners. Meanwhile, Cloud Logging provides real-time log management and analysis and can ingest custom log data from any source. The solution alerts Causaly on all of its log data and events, so that Causaly is able to use the information to quickly troubleshoot issues across its application.

Rapid, regular iterations are a key requirement for Causaly, which operates a monthly update cycle to its platform. Using scalable pay-as-you-go tools like Cloud Functions for running code with zero server management and Google Cloud Armor to protect applications against cyber threats, the company is able to scale its software efficiently and securely.

“There’s a lot of development in natural language processing and machine learning, and we are always trying to be competitive and understand where our technology is in terms of innovation, quality, and complexity,” says Saudabayev. “What we love about Google Cloud is that its ecosystem addresses many heterogeneous business requirements, from scaling capacity with Compute Engine to security with Google Cloud Armor, and much more. It’s the fact that we can build and iterate tools that can communicate with each other in one place.”

“The computing power and scalability of Google Cloud empowers us to continue developing our solution so we can help more researchers to make sense of all the information available to them worldwide.”

Artur Saudabayev, co-founder and CTO, Causaly

Tackling COVID-19 with intelligent network analysis

As Causaly has scaled, so has its ability to anticipate and react to the critical research needs of biomedical professionals. Nowhere was this clearer than in the company’s COVID-19 Network Analysis project, which mapped more than 60,000 full research papers on COVID-19 within six weeks from the start of the pandemic and made the results publicly available for noncommercial use. This was then released as one of Causaly’s monthly updates via App Engine.

“Already at the early stages of the pandemic, thousands of laboratories and scientists involved in this research area started publishing papers and developing different databases,” Saudabayev says. “We realized that this boom in research made it challenging for researchers to access a high-level view of all the information available. We set out to solve this problem by bringing together insights into all of this research on our platform.”

The project was testament to the ambition of Causaly’s growth story and its ability to keep scaling and meeting new challenges across the biomedicine research landscape.

“We're looking forward to growing together with Google Cloud and the initiatives that the company is taking,” says Saudabayev. “The computing power and scalability of Google Cloud empowers us to continue developing our solution so we can help more researchers to make sense of all the information available to them worldwide.”

Tell us your challenge. We're here to help.

Contact us

About Causaly

Causaly uses artificial intelligence to rapidly read, understand, and interpret vast databases of biomedical knowledge. Its platform surfaces evidence from 30 million biomedical publications in seconds, enabling researchers to rapidly map epidemiology data, biomarkers, genes, molecular targets, and identify potential treatment options.

Industries: Healthcare, Life Sciences
Location: United Kingdom