Micrograph of myeloma neoplasm from bone marrow biopsy
Investigators at GNS Healthcare and several research collaboration partners have published a study identifying the transcription factor PHF19 as a novel marker of aggressive multiple myeloma progression in newly-diagnosed patients.

Investigators at GNS Healthcare and several research collaboration partners have published a study identifying the transcription factor PHF19 as a novel marker of aggressive multiple myeloma progression in newly-diagnosed patients.

GNS Healthcare and partners said the identification of PHF19 could lead to improvements in how diagnostics developers target the transcription factor through single gene tests—as well as how researchers design clinical trials, and how oncologists make decisions on treating patients with newly diagnosed multiple myeloma.

“Using PHF19, a single gene test, could make the process of identifying the right patients for intervention or for clinical trial design faster and simpler,” Colin Hill, GNS Healthcare Chairman, CEO, and Co-Founder, said in a statement. “This challenge was a great display of collaboration and shows that together we can not only break down data silos but move toward identifying previously unknown regulators of disease with the help of AI.”

PHF emerged as a potential biomarker through the Multiple Myeloma DREAM Challenge, a crowdsourced effort to develop models of rapid progression in newly diagnosed myeloma patients, and to benchmark these against previously published models.

The model developed by GNS Healthcare and partner researchers consisted of four features: age, staging information and the expression of PHF19 and MMSET. Scientists validated the marker by conducting cell line experiments, through which they showed that knockdown of PHF19 led to reduced proliferation through cell cycle arrest in multiple myeloma.

Challenge participants were required to submit algorithms to identify prognostic factors marking disease progression from gene expression data.

Of more than 800 participants and nearly 200 algorithms submitted to the Challenge, the algorithm of GNS Healthcare was found to be most accurate in predicting markers of aggressive disease progression, outperforming even those available in commercial tests such as the gene expression programming (GEP) classifiers UAMS70 and EMC92.

Molecular datasets were contributed by GNS partners that included the Multiple Myeloma Research Foundation (MMRF), M2Gen, Dana Farber Cancer Institute, University of Arkansas for Medical Sciences, and the University of Heidelberg. The submissions included five microarray and three RNA-seq expression datasets that were annotated with clinical characteristics including gender, age, International Staging System and cytogenetics. The data was transformed to build models of multiple myeloma, using GNS Healthcare’s AI and simulation platform as well as Resilient File System (ReFS). ReFS is Microsoft’s newest file system designed to maximize data availability, scale efficiently to large data sets across diverse workloads, and provide data integrity by means of resiliency to corruption.

Researchers published their findings in the journal Leukemia. Fred K. Gruber, PhD, senior principal scientist at GNS Healthcare, was lead author of the study.

Gruber was joined by co-authors from Sage Bionetworks, University of Arkansas for Medical Sciences, Genome Institute of Singapore, Celgene Corporation (Bristol-Myers Squibb), Virginia Polytechnic Institute and State University, University of Michigan Medical School, Shanghai Jiao Tong University, Stanford University School of Medicine, M2Gen, Rancho BioSciences, Moffitt Cancer Center & Research Institute, Universitätsklinikum Heidelberg, Nationales Centrum für Tumorerkrankungen, Institut Universitaire du Cancer Oncopole, Dana-Farber Cancer Institute, Harvard TH Chan School of Public Health, Erasmus MC Cancer Institute, VA Boston Healthcare System, Multiple Myeloma Research Foundation, Labor für Myelomforschung, and Indiana University.

This site uses Akismet to reduce spam. Learn how your comment data is processed.