Σάββατο, 29 Ιουλίου 2017

WATSON USE TO ANALYZE GENOMIC DATA

A human–machine interface can significantly speed up the analysis of whole-genome sequencing, which in turn may address a "key bottleneck in cancer genomics," according to the authors of a new paper.
In a proof-of-concept study, published online in Neurology: Genetics, IBM's supercomputer Watson was able to interpret data much quicker than a team of experts.
The supercomputer  took 10 minutes to come up with conclusions that were similar to those reached by the team of experts after 160 hours of analysis.
The IBM Watson Genomic Analytics (WGA) is an automated system for prioritizing somatic variants and identifying drugs. In this study, it processed genomic data from a whole-genome sequencing and tumor RNA sequencing from a glioblastoma, as well as abstracts and full-text articles from PubMed.
The authors note that combining artificial intelligence with the whole-genome sequencing of tumor DNA could help clinicians pinpoint potential targets and therapies more efficiently. This is particularly important for aggressive cancers, such as glioblastoma, for which the 5-year survival rate peaks at 17%.
"Whole-genome sequencing is complex in terms of volume and increased amount of information," said study author Robert Darnell, MD, PhD, Heilbrunn Professor and Senior Physician at the Rockefeller University, New York City.  "This reaches into many areas of interpretation — from technical to biologic interpretation."
"To scale this, we envision a need for a human–machine collaboration to help teach — and ultimately training sets for artificial intelligence — machines to be able to do many of the aspects of scaling," he told Medscape Medical News.

Teaming Up With Watson

In the current study, Dr Darnell and a research team from the New York Genome Center, Rockefeller University, and IBM used a beta version of Watson for Genomics to interpret whole-genome sequencing from one patient with glioblastoma.
The patient had undergone an initial resection and was subsequently diagnosed with glioblastoma. After the procedure, he developed right-sided hemineglect and right/left confusion, became somnolent, and required another resection and ventriculoperitoneal shunt placement.
Two months later, he completed radiation therapy with concurrent temozolomide, and then completed three more cycles of adjuvant temozolomide at 5 months after the initial resection. His disease was progressive, and his oncologist recommended enrollment in a clinical trial targeting PIK3 and MET alterations. However, he declined rapidly, was no longer eligible for the trial, and died 8 months after the first surgery. This was shorter than the median survival time for this tumor type.
"This highlights one of the challenges of the clinical application of precision medicine technology," note the authors, emphasizing that the identification of targets and potentially useful drugs in a timely manner is only the first step.
"Drug and drug trial access is crucial to determine the benefit of this approach in cancer management," they write
In this case, the patient was helped "on paper," Dr. Darnell explained. "We identified a potentially relevant NIH [National Institutes of Health]-sponsored clinical trial with targeted therapies that matched the mutations/pathways seen with WGS [whole-genome sequencing] and RNA sequencing that were not otherwise visible," he said. "The referring physician was interested in pursuing this option."
However, the tumor progressed rapidly, despite treatment with standard therapy. "Time was a big variable in his outcome," Dr Darnell said. "This case therefore illustrates the potential and the challenge of where high-end cancer genomics is at and where it is headed."

Δεν υπάρχουν σχόλια: