The new mathematical algorithms could diagnose cancer
Dartmouth researchers (USA) have developed an algorithm that might someday be used to analyze blood for diagnostic purposes. Using data from a mass spectrometer, a device that generates a molecular fingerprint of biological samples, equipment Dartmouth calculations can distinguish healthy blood from diseased blood. This study by Ryan Lilien, a Dartmouth medical / doctoral student, Hany Farid, Assistant Professor of Computer Science, and Bruce Donald, professor of computer Foley, published in the Journal of Computational Biology in December 2003.Our algorithm called Q5, assumes that the molecular composition of blood changes between healthy and diseased state, says Donald, the senior researcher on the project. The aim of our work, minimally invasive diagnostic procedures
should be developed with high predictive power, and this is a promising first step. mathematical calculations are routinely developed, varied and refined to analyze mass spectrometry. Q5 uses mathematical techniques called Principal Component Analysis (PCA) and linear discriminant analysis to distinguish (LDA) to switch between the mass spectra of blood samples from healthy and sick.Q5 learns with each sample for testing, which to a greater accuracy. The algorithm compares the molecular fingerprint of each sample, the characteristics that identify which vary from state health and disease. Our algorithm for detecting ovarian cancer with almost 100% accuracy and prostate cancer by about 95% accuracy, said Lily, the main author of the article. Q5 analyzes data from mass
spectrometry and the threshold between healthy and disease classification control. Though we are against ovarian and prostate cancer test, we think its possible that Q5 be used to test for other cancers and diseases.The researchers said that there was still much to learn different types of information within a blood sample, and Q5 is a means of gaining new and important data. More exciting for us, in contrast to previous
Pages: [1] 2