Broadly accepted good practice standards for the development of prognostic markers and prognostic classification systems do not exist (Simon and Altman, 1994).
Both clinicopathological markers and molecular classification have clearly indicated that breast cancer is a heterogeneous disease with distinct clinical outcomes [ 2, 31].
In colon cancer data used for marker computation and classification test data come from two different platforms.
Recently, it has been suggested that fusion status may be a better stratification marker than histology, and classification of RMS into fusion-positive versus fusion-negative (rather than ARMS and ERMS, respectively) may be more useful in prognostication and clinical allocation of therapy3.
We have also examined individual mitochondrial genes (cox1, cob, nad5) and rDNA internal-transcribed spacers (ITS), which have been proposed as markers for species identification and classification to complement common nuclear DNA markers [ 10].
The results obtained are discussed with available data from nuclear markers and with morphological classifications, and all this information is analysed in relation to the origin of the different Iberian sheep breeds.
In considering the functional utility of the maxicircle as a molecular marker for strain identification and classification, the coding regions present variation that is likely conserved among DTUs or subgroups of strains, while the variable regions may allow differentiation among closely related isolates.
Over the past decades, an increasing amount of evidence has demonstrated applications of microRNAs (miRNAs) as tissue based markers for classification and prognosis of several human cancers, including breast cancer [ 137, 138].
The proposed two-step Bayes classification framework was equal to and, in some cases, outperformed other classification methods in terms of prediction accuracy, minimum number of classification markers, and computational time.
For breast cancer TP53 wildtype versus mutant markers are computed using GSE3494 and classification is performed by leave-one-out cross-validation in the same dataset.
