I have long been fascinated with data’s power to drive new insights, including developing a deeper understanding of human physiology/pathophysiology, informing decision-making, and guiding the development of new treatment options. Often novel complex analysis methods are linked to specialized visualizations to drive the development and use of these insights. My fascination has grown into a career that aims to decrease the time required to go from data to insight.
I recently presented a webinar titled, Ushering a New Era of Precision Medicine with Muti-Omics, I reviewed several real-world use cases, many I have had the privilege to work on at Velsera. Here, I elaborate on some of the points discussed in the webinar.
The Era of Multi-omics is happening now and Velsera is already supporting it. In the webinar, I shared use cases that leverage a number of Velsera-powered platforms, including the development of multi-agent treatment options for cancer on the Cancer Genomics Cloud, use of proteomics data to identify Cardiovascular phenotypes for Black adults using the NHLBI BioData Catalyst Platform, and multi-modal analysis of pediatric brain cancer on CAVATICA. The features of these use cases include multi-omics data, multi-modal data including imaging, and harmonization with established clinical phenotypes. These use cases are not an anomaly, they are just a few examples of multi-omics and multi-modal research programs that are increasing every quarter over the last three to four years. The increase in multi-omics-modal research is now a frequently seen hallmark of life science research leading to advances in health care.
Scaling multi-omics capability is required to meet the rapid increase of multi-omics precision medicine programs. Translating complex analysis into a form that can process data with reliability, and reproducibility is the challenge of every analyst that has ever worked in the life sciences. The new era of multi-omics adds a few new requirements. We are seeing increases in the size of production-grade multi-omics workflows, with workflows having dozens of tools and nested workflows linked that can easily involve hundreds of parameters. Workflow languages and managers are essential. One way Velsera increases the scale of our computational platforms is to build our workflows with the Common Workflow Language (CWL). The advantage is that the workflow is portable and can be used across many computational ecosystems. Moreover, our systems can use the information within CWL to automatically scale data analysis (ex., Batch processing of data inputs and scheduling required resources) on the cloud without user input. We have also seen substantial cloud cost saving by using run information to optimize workflow execution to decrease analysis times. We also publish standardized workflows in the Seven Bridges Public Apps Gallery, which are optimized to support multi-omics analysis, including whole exome, RNAseq, proteomics, and metabolomics. These workflows made available on the platform allow our partners to start standard analysis nearly immediately and provide a stable starting point for novel analyses. In addition to providing world-class, optimized CWL tools and workflows out of the box, Velsera platforms also support other descriptive workflow languages, including Nextflow and WDL, so that developers can port in existing tools and leverage them alongside public apps workflows, reducing the need to reinvent the wheel. Removing the technical barriers required to scale multi-omics analysis allows teams to focus on generating project outcomes, including interpreting analysis results.
Meeting healthcare objectives requires collaborating across scientific, technical, and analytical challenges. My team is often called to advise on complex analytical analysis that uses large datasets and complex analysis methods, which increasingly utilize machine learning techniques. I have found that projects that move forward the fastest and result in better outcomes are carried out by teams that include cross-disciplinary members committed to achieving outcomes. Consequently, supporting a project might include engaging with a specialized team that pulls from Velsera bioinformatics, partnership, and engagement and training teams. Establishing a vision and fostering supportive interactions to help researchers achieve proficiency and independence, are catalysts for establishing the long-term success and adoption of computational platforms in a research setting. Consequently, our team at Velsera works closely to bring our partners from novices to expert users by providing high-quality documentation and interactive and online training materials.
We need diverse communities across the life sciences to collaborate within a connected ecosystem that reduces the technical burdens required to conduct multi-omics research programs. For over 10 years, Velsera has invested in collaborative computational platforms that allow scientists, physicians, researchers, and bioinformaticists with varying levels of skill and domain expertise to collaborate effectively without the constraints posed by geography and affiliation. This provides a platform to bring together communities of interest to collectively address complex scientific questions. We invite you to partner with us to solve the most pressing problems challenging health care today.
The importance and need for diversity in data sets cannot be underestimated as we explore answers with a multi-omics lens. There is a growing need for greater degree of racial and ancestral diversity in multi-modal data sets to enable us to meet healthcare needs of patients from diverse social and ethnic backgrounds.
Want to learn more?
Are you interested in learning more about the power of multi-omics and multi-modal data? Register for our upcoming webinar, “Advancing Precision Medicine: Multi-Modal Analysis and Imaging for Enhanced Diagnosis and Treatment.” During this session, we will discuss the growing significance of imaging in precision medicine applications and discuss the transition towards integrating machine learning and artificial intelligence (ML-AI) techniques in multi-modal analysis in the cloud.