Fig 4.tif (526.98 kB)

Receiver Operating Characteristic (ROC) curve and Decision tree.

Download (526.98 kB)
posted on 01.04.2016, 06:53 by Elodie Granjon, Marie-Laure Dichtel-Danjoy, Esber Saba, Ester Sabino, Lea Campos de Oliveira, Maan Zrein

(A) The graph is showing the ROC curve analysis of the sensitivity and specificity of each of the 12 antigens. IBAG36-IBAG37 and IBAG95 gave the bests results. (B) The decision tree established after the first data analysis is based on the number of reactive antigens. A sample is considered as negative if reactive on up to 3 antigens and positive if reactive on at least 6 antigens. When reactive on 4 or 5 antigens the sample is further analyzed: the Sum of intensities on three antigens (IBAG36-37-95) is measured and the 12 000 raw value is used as a threshold.