Alex Philippidis Senior News Editor Genetic Engineering & Biotechnology News

Computing tools may finally turn personalized-medicine hype into hope.

Personalized medicine has long offered more hype than hope. But as genetic knowledge has multiplied in recent years, researchers—and more importantly, computing tools—have begun catching up with all that far-flung data, harnessing it into new databases and systems that offer the best prospects yet for delivering on the promise of precision treatments.

In several papers published this year for which he was corresponding author, Leo Anthony Celi, M.D., M.P.H., clinical research director for MIT’s Laboratory of Computational Physiology, joined David J. Stone, M.D., a visiting professor at University of Virginia and faculty associate at UVA Center for Wireless Health, and others in discussing the challenges that can be addressed with new computing systems, and the key data such systems must capture for clinicians.

In June, Drs. Celi and Stone joined Andrew J. Zimolzak, M.D., a research fellow at Children’s Hospital Boston, in proposing an operational vision for real-time incorporation of external health data through “dynamic clinical data mining” (DCDM), which they envision as driving next-generation electronic medical records (EMRs) as well as “turning medical practice into a data-driven, logical, and optimized system.”

“The next step in the clinical digitization process should be the creation of a medical Internet equivalent that incorporates the rapid, powerful data search engine features that all current Web users employ,” the researchers wrote in the Journal of Medical Internet Research (JMIR) Medical Informatics.

DCDM would underpin a system where data on the content and context of an individual patient’s care encounter would be rapidly compared and matched with those of a population database—complete with hyperlinks to detailed data, conventional practice guidelines, and evidence-based modalities.

“The system would identify and suggest prioritized interventions and other courses of action that have been shown to be most valuable in terms of outcome and cost,” the researchers explained.

Mimicking MIMIC

They suggested as a prototype the publicly available Multiparameter Intelligent Monitoring in Intensive Care (MIMIC) II Database. MIMIC-II includes information from EMRs of patients admitted to the intensive care units at Beth Israel Deaconess Medical Center since 2001. More than 1,000 investigators from over 32 countries have unrestricted access to deidentified clinical data via MIMIC-II.

“The ideal data system would also be organized in modules representing particular organs or disease states with these modules nested in the global dataset reflecting system pathophysiology,” Drs. Celi, Stone, and Marie Csete, M.D., Ph.D., chief scientific officer at Huntington Medical Research Institutes, wrote in the October issue of Current Opinion in Critical Care.

Dr. Celi, an assistant clinical professor at Beth Israel Deaconess Medical Center, told Clinical OMICs that MIMIC-II’s biggest challenge has been the capture of outcome data. While EMRs have noted whether patients got discharged or not, they do not have data on how long those patients survived. MIMIC-II has sought to fill that gap with data from the Social Security Death Index.

The investigators had to match outcome data to see whether the patient was still alive or had died, in which case post-discharge survival time had to be determined, said Dr. Celi, who is also founder and executive director of Sana, a nonprofit created to improve quality of care in resource-poor settings. “Clearly, mortality is not the only important outcome that we would like,” added Dr. Celi. “Quality of life is always the best outcome that we would like to capture. That’s almost impossible, but there are some proxies for quality of life.”

Drs. Celi, Stone, and colleagues found a significant increase in hospital-stay mortality among intensive care unit patients who took selective serotonin reuptake inhibitors or serotonin norepinephrine reuptake inhibitors before admission, compared with a control group. The risk was highest in patients with acute coronary syndrome and in patients admitted to the cardiac surgery recovery unit, the researchers reported in the journal Chest.

In the journal Big Data, published by Clinical OMICs publisher Mary Ann Liebert Inc., Drs. Celi, Stone, and colleagues published an article (“From Pharmacovigilance to Clinical Care Optimization”) that proposed an open database on delayed and low-frequency adverse events connected with new drugs.

CoMMpass Bearing

Developing software systems for precision medicine has been the focus of Genospace, a Cambridge, MA company that creates systems designed to securely store genomic and health data for retrieval in formats specific to clinical practitioners, laboratories, researchers, and patient communities. The company’s offerings can serve any combination of users depending on specific needs.

In addition to curating variants and molecular associations from clinical trials, GenoSpace integrates reference content from multiple sources—including Thomson Reuters’ Cortelis platform and public databases such as the National Center for Biotechnology Information’s ClinVar, the National Heart, Lung, and Blood Institute (NHLBI) Grand Opportunity Exome Sequencing Project, the Thousand Genomes Project, and the European Bioinformatics Institute, part of the European Molecular Biology Laboratory (EMBL-EBI).

GenoSpace’s chief operating officer Mick Correll told Clinical OMICs that the company’s platform not only enables pathologists, medical directors, and other users to create a unified report summarizing data across multiple technologies, it also helps clinicians answer key questions as they develop treatment plans by accessing therapeutic and prognostic data and data from clinical trials relevant to a patient’s case.

“To make a dataset like that previously required research funding,” Correll said. “The pathology labs are really seeking to combine high-dimensional results with the information coming from more traditional assays. Sometimes, it can be a single-marker PCR test with immunohistochemistry, with cytogenetics and FISH and flow cytometry, and complementing that with a next-gen sequencing panel, and copy number alteration assessment.”

Earlier this year, GenoSpace joined Thomson Reuters and pathology services provider PathGroup to expand the SmartGenomics™ Heme Profile from solid tumors to blood cancers. The profile provides genomic profiling and information for patients that have failed or been unsuccessful on repeated therapy for the group of cancers that include leukemias, lymphomas, and myeloproliferative/myelodysplastic diseases.

The Heme Profile includes next-generation DNA sequencing of 83 clinically actionable genes mutated in hematolymphoid cancers. The Solid Tumor panel now stands at 38 genes—with six genes added and three deleted since launch. Later this year, more customized gene panels for specific disease areas are set to be introduced, Correll said.

GenoSpace also partners with the Multiple Myeloma Research Foundation (MMRF) in a publicly shared database with portals or “gateways” allowing researchers to access data from MMRF’s Relating Clinical Outcomes in Multiple Myeloma to Personal Assessment of Genetic Profiles Study (CoMMpass), while also letting patients access real-time clinical and community support.

More than 1,000 patients have been screened at 92 sites worldwide, with 669 patients enrolling in CoMMpass, toward a goal of 1,000. The most recent release to the research gateway contains clinical information on 626 people and complete genomic information on 363. The 10-year study has produced the most complete multiple myeloma patient dataset available worldwide for collaborative research.

Beyond Proteins

In September, a consortium of U.S., European, and Asian universities and research institutions launched RNAcentral, the first public database envisioned as a unified resource for noncoding RNA data. The first release of RNAcentral contained about 8 million sequences; as of September 12, it had integrated 10 of the consortium’s 25 databases: European Nucleotide Archive; Rfam; RefSeq; VEGA; gtRNAdb, RDP; miRBase, tmRNA Website, SRPDB, and lncRNAdb.

While it’s still early, RNAcentral could help efforts to develop precision therapeutics based on noncoding RNAs—and help realize the longtime promise that personalized medicines could be developed from RNA after years of logistical hurdles such as delivery of the fragile genetic material to target cells.

“Currently, people are only really looking at the protein-coding genes when they look at these RNA-seq experiments,” Alex Bateman, EMBL-EBI’s head of Protein Sequence Resources, told Clinical OMICs. “They tend to do very little on looking at the RNA levels for noncoding RNAs. RNAcentral is really going to enable us to use all types of macromolecules, not just the proteins.”

This article was originally published in the October 8 issue of Clinical OMICs. For more content like this and details on how to get a free subscription to this digital publication, go to

Previous articleCSL Behring Plans $450M Expansion of Illinois, Australia Sites
Next articleEndo Scores with $2.6B Offer for Auxilium, Which Ends QLT Deal