HypaHub
A Guide to Faster and Better Bioinformatics

Since the debut of next generation sequencing, biomedical data is being generated exponentially, with ~2 to 40 billion gigabytes of data generated yearly. This vast collection of DNA data can improve understanding of diseases and unlock new personalized treatments. HypaHub sits at the intersection of statistics, computer science, and biology to facilitate what is currently most challenging and demanded: diagnosing and curing diseases with data. The company aims to democratize bioinformatics and make it accessible for everyone, maximizing its impact in facilitating biomedicine.

HypaHub is a cloud-based bioinformatics and AI platform. It is the first HIPAA-compliant bioinformatics SaaS that runs natively on users clouds, providing both operation and price transparency. The company's product is a one-stop shop to access complex technologies designed for scientists building data applications or performing data analytics to extract insights.

"Be it a lab scientist or data scientist, our platform offers them a tool to analyze data at scale, such that they can build, manage, and deliver biomedical analytics faster than ever," states Hugo Lam, founder and CEO of HypaHub. Hugo has over 20 years of experience in software, bioinformatics, and genomics, including time at Roche, Personalis, and 23andMe. Before founding HypaHub, he was senior director and head of bioinformatics at Roche. He is an alumnus of HKUST, Imperial College London, Yale University, Stanford Medical School, and Cambridge Judge Business School. He has published over 40 publications with more than 35,000 citations.

Why HypaHub?

Skilled professionals and cutting-edge technologies are vital to the success of diagnostics and pharmaceutical companies. However, many complex technologies, heterogeneous data, and algorithms in the genomic domain are not easily accessible. For example, multi-omics analysis, active learning, deep learning, single cell sequencing, or even quantum computing, to name a few. HypaHub effectively encapsulates the complexities of these technologies, making them accessible to those with various levels of experience and expertise. HypaHub integrates with clients' toolchains, and with a few clicks, users can launch complex workflows with terabytes of data or perform interactive analyses for insight.

A biomedical project always involves people with different domains of expertise, such as project managers, lab scientists, data scientists, software engineers, and sometimes external collaborators. However, there are often unintended silos due to the lack of a platform that can facilitate collaboration. HypaHub allows various departments within an organization to join forces on a project and share data and applications.

Due to a data center's high operation and maintenance costs, the cloud is a no-brainer for most, if not all small and medium-sized enterprises. Moving to the cloud, though, is more expensive than they initially think. The cost to avail cloud computing with the capacity of a MacBook can range from $2k to $6k over an entire year.

In a biotech organization, resources like that are oftentimes idle 50% to 70% of the time. HypaHub allows users to set a time frame, usually outside their work hours, for cloud computers to intelligently sleep when detected as idle. This can save up to 70% of the cost. In conjunction with their technology that leverages the use of object storage, it will also lower the storage cost by almost 80%.

Be it a lab scientist or data scientist, our platform offers them a tool to analyze data at scale, such that they can build, manage, and deliver biomedical analytics faster than ever

Another common use that incurs cost is batch processing, a method of running software programs in batches automatically. A cancer cohort can easily consume many hours on hundreds of computers. Even though these computers could be spot instances on the cloud, potentially lowering the cost, the downside of using spot instances is that they can be preempted at any time during computing. HypaHub keeps all the intermediate data, so a complex computation workload does not have to start from scratch whenever it fails. This alone can save up to 90% of the cost compared to using on-demand instances, to achieve the exact same computation.

In the era of computational biology, enterprises need to hire software engineers to develop software and algorithms set to their technologies. Highly skilled professionals have to divide their attention between building infrastructure code, deploying user tools, and developing bioinformatics workflows, before they can focus on delivering what they are keen to do. At the same time, maintenance occupies about 30% of their time, some studies found. HypaHub solves this problem by handling all the tedious work and heavy lifting by automating the creation and management of computing resources, eliminating the need for enterprises to set up and maintain a high-end R&D cloud infrastructure and an associated technology stack.

HypaHub enables clients to focus on their core business and lower their spending on IT. HypaHub orchestrates resources for their bioinformatics needs while clients retain ownership of the resources and are in complete control of the cost. For instance, a precision medicine company is utilizing HypaHub to orchestrate their computational resources and generate bioinformatic insights leading to accelerated scientific discovery, while minimizing costs. Another customer, a diagnostic company, is also availing HypaHub to automate their RNA analysis.

What Else is There?

HypaHub incorporates professional applications such as HypaBfx, HypaML, and HypaCluster. HypaBfx is a high-quality software toolkit that contains various popular and robust bioinformatics tools. HypaCluster is a parallel cluster providing a virtual, scalable, and dynamically provisioned cluster for batch computing, unlike traditional clusters that require costly static resources. It allows users to submit jobs to an auto-scaling cluster in an interactive environment, like a command prompt or portable batch system (PBS), which are dear to many in the supercomputing world. HypaHub has also developed HypaML, a machine learning module.

It is a statistical and machine learning application that enables data scientists and data explorers to quickly extract insights from their big data. It provides an intelligent interface to facilitate analysis of biomedical and genomic data. With a single click, the data models generated can be easily exported to a Jupyter Notebook for further customization, without leaving the platform.

A Cut Above the Rest

Unlike many bioinformatics products focused on a single feature, HypaHub provides a comprehensive, end-to-end platform to serve clients’ daily bioinformatics needs. HypaHub provides various applications within its platform through an app store, where users can also develop and publish their applications. Organizations usually have their proprietary information and algorithms, which they can share with other users within their company through HypaHub. The platform automates the allocation and scalability of resources on different genomic applications while it monitors their life cycles, like shutting them down or reporting issues. HypaHub delivers a fully dedicated and isolated infrastructure provisioned under a user's cloud account for complete control and ownership of their resources.

HypaHub boasts a team of professional engineers and bioinformatics scientists. The company can understand today's bioinformatics needs and develop a platform with high-quality software so users can perform end-to-end, routine bioinformatics daily. From diagnostics to pharmaceuticals, the company works closely with clients and accepts feature requests to facilitate their operations.

HypaHub's user-friendly, collaborative platform is the first to run natively on a user's cloud. From subscription to computation, the firm maintains transparency, and there are no markups on cloud resource costs. The platform can be utilized by both enterprises and academia. With work done via HyapHub published in a number of scientific journals and conferences, it has proven to be a robust platform for scientific discovery and insights.

What Lies Ahead?

Innovation is the engine of HypaHub. Based on clients' and its experts' feedback, the firm is planning to add more modules to the platform. Since genomics is comprehensive by nature, HypaHub is ready to leverage cutting-edge, complex technologies to facilitate its clients in extracting insights from enormous data. From a resource perspective, HypaHub is focused on developing analytical tools and apps for users and providing more valuable data. For non-developers, HypaHub will continue to simplify the user experience so they can achieve more. For developers, the company will provide more APIs and SDKs to integrate with their existing systems. Based on its highly scalable, HIPAA-compliant, and multi-modal platform, HypaHub is currently developing solutions with advanced AI to facilitate the most demanding pharmaceutical and clinical needs, including biomarker and drug discoveries.