A Molecular Examination of Benign and Malignant Melanocytic Tumors by Histology-Directed Tissue Profiling

摘要

1. Genetic Algorithm Classification model created with a teaching set of 20 samples for each invasive malignant melanocytes and benign nevi. 2. Classification model consisting of 15 peaks. 3. Leave 20% out cross validation results in an overall 93.63% classification accuracy. 4. Two of the peaks from the classification model identified as Vimentin (m/z 1428) and Actin (m/z 1198). 5. A more sensitive way to analyze melanocytic samples that are not visually discernible.