In MRMR, the mean value of all mutual information values across individual features and class labels are calculated as the max-relevance values.
To achieve this, the percentage of raters' labels (excluding abstentions) agreeing with the original author-based labels was calculated, both overall and for each item separately.
Relative quantification of proteins using the TMT labels was calculated using the tandem mass spectrometry scans as the ratio of the areas under the peaks at 126, 127, 128, 129, 130 and 131Da, which correspond to the representative masses of the TMT reagents.
The accuracy rates of red patch labeling were calculated by comparing the detected regions with the GT ones.
DNA adduct levels (RAL, relative adduct labelling) were calculated from the adduct cpm, the specific activity of [γ-P]ATP and the amount of DNA (pmol of DNA-P) used.
After chromatography TLC plates were scanned using a Packard Instant Imager (Dowers Grove, IL, USA) and DNA adduct levels (RAL, relative adduct labelling) were calculated from the adduct cpm, the specific activity of [γ-P]ATP and the amount of DNA (pmol of DNA-P) used.
After chromatography, TLC sheets were scanned using a Packard Instant Imager (Dowers Grove, IL) and DNA adduct levels (RAL, relative adduct labeling) were calculated from adduct cpm, the specific activity of [γ-P]ATP and the amount of DNA (pmol of DNA-P) used.
For each iteration, new pseudo-labels are calculated for the unlabeled target instances and the process of mapping, classifying, and updating continues until convergence is reached.
Two classifiers are built for two newly constructed views for the target data (denoted as f 1 and f 2) where f 1 is initialized to the source prediction function h and f 2 is initialized to 0. The predicted label is calculated as a function of these two views thus creating an ensemble classifier.
Degree of labeling was calculated to 3 mole dye per mole protein.
