High Performance Computing Solution to Overcome Current Limitations in Mass Spectrometry

Jeff Peterson, Chairman at Veritomyx and CEO of parent company, Target Discovery, talked about how Veritomyx is focusing on convergence of high performance computing and state-of-the-art signal processing algorithms, to deal with the challenges encountered in identifying protein/isoform and metabolite biomarkers, due to limitations in mass spectrometry and supporting software.  The talk was in Sunnyvale, at event hosted by www.bio2devicegroup.org .

Modifications of proteins are responsible for exerting about 90% of the control of biochemical pathway changes. Accurate detection and validation of these changes offers hope for much stronger diagnostic insights into disease status and valuable treatment guidance decisions, enabling new effectiveness to advance the field of personalized medicine.  In turn, this is expected to help determine specific potential responses of therapeutic compounds in individual patients, allowing for more customized and successful treatment regimens.  Selection of properly validated therapies would save huge amounts of resources, money, and most importantly improve patient outcomes.

In the US, healthcare expenditure currently averages around $3T per year, said Peterson.  About 85% of that is spent on doctors, hospitals, ER, and end of life management.  Drugs make up only 12-15% of the total expenditure and diagnostics represent 2-3%.  The dawn of personalized medicine diagnostics now allows the right diagnostic to better select for the right therapeutic, and to minimize the patient’s descent into the chasm of other healthcare expenditures.  This paradigm shift is upending the value proposition in favor of newly recognized value in personalized medicine diagnostics leveraging overall healthcare expenditure decreases with improved outcomes.

A Shimadzu Ion Trap-Time of Flight mass spectrometer (center), HPLC-UV (left). (Photo credit: Wikipedia)

The search to find, identify and characterize new biomarkers across life science R&D efforts typically employs mass spectrometry.  In MS, ions from the unknown are accelerated and detected electro-magnetically, and software attempts to assemble puzzles of component ions into information identifying the unknown parent.  The output is typically analyzed in the form of a diagram, with X axis reflecting the masses of the individual ions, and the Y axis showing the  “relative abundance” of the ions detected.

 

However, current limitations in mass spectrometry, signal processing and identification algorithms often limit the availability of useful information.  Frequently, critical information remains hidden among overlapping peaks.  Overlapping peaks cause misdirection in mass and abundance results, delivering lower mass and abundance precision, and making it vastly harder or completely failing to correctly identify biomarker compositions.

English: Isotopic pattern of an peptide. Mass spectrum recorded by Q-TOF mass spectrometer. Polski: Obwiednia izotopowa peptydu zmierzona przy pomocy spektrometru Mas Q-TOF (Photo credit: Wikipedia)

The “PeakInvestigator™” software from Veritomyx, helps enhance the quality of science, saving critical time and R&D resources.  PeakInvestigator typically triples the effective resolution and reveals and precisely deconvolves overlapping peaks that go unnoticed by current mass analyzers and software.

Precision problems in the data not only waste a great deal of time, but they can misdirect attention and decrease the effectiveness of R&D efforts.  Abundance and mass errors produce spurious biomarker candidates and incorrect isotopic ratios for peak identification, said Peterson.  The PeakInvestigator software delivers up to 10X improvements in mass and abundance precision across a wide dynamic range when deconvolving previously hidden overlapping peaks.

The PeakInvestigator is based on statistically robust and reproducible methodologies, said Peterson and it enhances identification of unknown proteins, peptides, and metabolites to improve the ability to correctly detect and validate biomarkers.  Further, it’s advanced signal processing algorithms automatically adjust for differences in instruments and tuning variations.  These algorithms provide adaptive baselining and signal-to-noise thresholding.  This easy to use software solution eliminates the need for manual estimation of centroiding parameters or visual inspection of data before processing.  When summing up the advantages, Peterson also added that PeakInvestigator is being designed to be 21CFR11 and HIPAA compliant.

For companies interested in PeakInvestigator software-as-a-service solution, it can be easily integrated into the existing workflow.  Their simple public application-programming interface (API) inserts directly into company’s existing mass spectrometry workflow.  While already validated workflow remains intact, the PeakInvestigator algorithms can be leveraged on the existing data.  Currently Veritomyx has launched Free Beta Services collaborations for its PeakInvestigator deconvolution and centroiding software.  Free Beta Services will allow collaborators to apply certain PeakInvestigator high performance computing capabilities to their own mass spectral raw datasets at no charge.   Please contact www.veritomyx.com for further information.