Intel Powers Precision Medicine Technology
Precision medicine delivers personalized treatments and patient-centric care for specific subsets of individuals that share genetic markers, traits, or conditions. The technology required to enable this transformative treatment model spans from data generated at the edge to data science workloads on High Performance Computing (HPC) platforms in the cloud or on-premises. Clinical applications, academic research, and pharmaceutical discovery and manufacturing require powerful instrumentation, analytics, and HPC.
As part of our mission to create world-changing technology that improves the life of every person on the planet, Intel plays an integral role in enabling precision medicine solutions. Manufacturers use our hardware technologies and software tools to create precision medicine solutions such as those that capture biological processes, analyze biomarkers, and predict drug interactions.
Whether we’re collaborating with an AI-enabled lab equipment provider to optimize edge performance on next-gen sequencers, working with an OEM or cloud provider to build HPC architectures that accelerate advanced analytics, or helping an ISV to optimize their solutions on Intel® architecture, Intel is dedicated to improving lives through genomics and precision medicine technology.
For example, we’ve worked with the Broad Institute to create an Intel® Select Solution that helps organizations deploy scalable HPC clusters for genomic analytics. The latest release enables users to run the Broad Institute’s GATK Best Practices Pipeline for Germline Variant Calling to process up to eight whole genome samples per compute node per day.
We’ve also worked with many health and life sciences application providers to optimize their code for improved performance and an enhanced user/developer experience.
You can find more examples of the innovations we’re enabling in the Intel® Solutions Marketplace.
Precision Medicine Tools and Techniques
Precision medicine uses techniques such as molecular diagnostics (which includes genetic testing), molecular imaging, next-generation sequencing, and molecular dynamics to diagnose disease and tailor treatments to the individual.
Molecular Diagnostics and Genomic Analysis
Molecular diagnostics involves analyzing a patient’s biomarkers through genome sequencing. These genetic tests reveal information that can be used to provide the most effective treatment or predict which drugs will work best for the patient.
Advancements such as next-generation sequencing have dramatically impacted genomic research. While traditional sequencing technology required over a decade to deliver results, next-generation technologies enable an entire human genome to be sequenced in approximately 40 hours.
Molecular imaging, specifically a method called cryogenic electron microscopy (or cryo-EM), is used extensively in the drug discovery process to provide insights into the 3D structures of proteins and other biological entities. More recently, this technology has also been applied to identify small drug-like molecules. Other forms of molecular imaging fall under medical imaging, including the use of various methods (MRI, CT, PET) with a contrast agent to visualize, characterize and quantify biological processes in a patient.
Molecular dynamics is a computational method that estimates the possible three-dimensional configurations of chemical and biological structures as well as simulates their motion. Molecular dynamics is often used in conjunction with molecular docking, often referred to simply as “docking.”
Docking is a method used to predict how effectively small molecule drug candidates will interact with a protein of interest for a particular disease. These simulation methods enable the assessment of a binding interaction for each drug candidate and ultimately allow researchers to create a ranking for the effectiveness of the various drug candidates.
An End-to-End Platform for Genomics and Precision Medicine
From a technology perspective, precision medicine involves both medical devices or PC workstations at the edge and IT systems in the data center. Intel® hardware powers solutions across that continuum.
For example, in genomics, our edge technologies such as Intel Atom®, Intel® Core™, and Intel® Xeon® E processors enable a rapid primary analysis of samples in the lab equipment offered by many Intel partners. Intel® Xeon® Scalable processors also fuel the base-calling algorithm and sequence alignment tools used for secondary analysis in back-end systems. Throughout all these examples, our memory, GPU, storage, and connectivity technologies can also be used to increase throughput and overall solution performance.
High Performance Computing in Precision Medicine
For the secondary analysis, HPC systems distribute these intensive computational workloads across several centrally managed nodes, processing information in parallel to dramatically accelerate results. Intel-based HPC infrastructure helps enable faster discoveries and more-effective research to improve predictions and models.
As a critical part of many HPC systems supporting precision medicine workloads, Intel® Xeon® Scalable processors help facilitate progress for key use cases and common applications such as:
- Genomics: Improving our understanding of individuals’ genetics for better health outcomes with applications such as the Genome Analysis Toolkit (GATK) 4.x.
- Cryo-EM: Determining the molecular structures for biological studies and drug development with RELION 3.x.
- Quantum mechanics: Describing the interactions between atoms, molecules, and macromolecules with VASP and NWChem.
- Molecular dynamics: Simulating and analyzing the movements of atoms and molecules with NAMD, GROMACS, and LAMMPS.
While each of these research-critical workloads has different requirements, they all share a similar demand for intensive computing power on a large scale. 3rd Gen Intel® Xeon® Scalable processors support this need with accelerated throughput alongside capacity for more I/O and memory. We work closely with partners throughout the industry to help research organizations, solution providers, and software vendors select and tune the right hardware for each workload.
Building these compute-intensive applications for use on distributed HPC systems presents unique software development challenges. Developers can build, analyze, optimize, and scale HPC applications across multiple types of architectures more easily using the Intel® oneAPI Base Toolkit and Intel® oneAPI HPC Toolkit. These resources include state-of-the-art techniques in vectorization, multithreading, multinode parallelization, and memory optimization.
Artificial Intelligence (AI) in Precision Medicine
Increasingly, precision medicine tools are being infused with AI capabilities to accelerate research that leads to positive health outcomes. These workloads can take place on distributed HPC systems or at the edge, where AI models are embedded into lab equipment to help expedite processes and empower data scientists and researchers to uncover insights more quickly and efficiently.
At Intel, we provide the accelerated AI computing that researchers, device manufacturers, and software providers need for rapid, seamless performance. Intel® Xeon® Scalable processors deliver built-in AI acceleration via their Intel® AVX-512 and Intel® Deep Learning Boost features, providing an easy way to supercharge AI performance in the data center.
In the lab, we help device manufacturers and software developers to more easily incorporate AI into their solutions. While AI has typically been thought of as requiring expensive, specialized hardware, Intel® tools and innovations are making it possible to integrate advanced AI algorithms directly into laboratory equipment using affordable general-purpose compute hardware.
For example, the Intel® Distribution of OpenVINO™ toolkit helps simplify and optimize training and deployment of AI inferencing algorithms across architectures, making it easier to deploy familiar frameworks on cost-effective Intel® architectures without sacrificing performance or accuracy.
To help accelerate AI and analytics on HPC infrastructures, Intel offers the Intel® oneAPI AI Analytics toolkit. This comprehensive package provides data scientists, AI developers, and researchers with familiar Python tools and AI frameworks to accelerate end-to-end data science and analytics pipelines on Intel® architectures.
Like the Intel® oneAPI HPC Toolkit, the Intel® oneAPI AI Analytics Toolkit’s components are built using oneAPI libraries for low-level compute optimizations. The toolkit maximizes performance end to end—from preprocessing through machine learning—and provides interoperability for efficient model development.
A critical component to creating AI precision medicine solutions is bringing together a variety of data sources securely to better train AI algorithms and models. Here, Intel® Software Guard Extensions (SGX) can be used to enable protected federated learning—protecting workloads and data in secure confidential computing enclaves, in the cloud or on-premises, to maintain privacy and security standards.
Genomics Optimization for Cloud Analytics
Intel is focused on enabling genomics workloads to run with optimal performance and flexibility in the cloud. We’re working to foster cloud-native designs among our ecosystem partners so that customers can experience positive results as they move to the cloud, with benefits across both cost and performance.
Additionally, our efforts with the Broad Institute have included optimization of GATK Best Practices hardware recommendations for genomic workloads across on-premises, public cloud, and hybrid cloud use cases.
Pioneering the Future of Precision Medicine
As precision medicine continues to evolve, Intel is committed to working with partners, researchers, and healthcare providers to continue driving innovation and advancement. From both a hardware and software perspective, we provide the essential elements required to build precision medicine technology and genomics while simplifying workflows, accelerating insights, and enhancing care for patients across the globe.