Skip to content

Scoring methods

Julian Uszkoreit edited this page Mar 7, 2014 · 2 revisions

Whether a higher or a lower protein score is better, depends directly on the scoring method. For "Additive Scoring", a higher score is assumed to be better. For both other scorings, a lower score is better. Note, that the protein score is not the product of the PSM scores, but the sum of the logs for numerical safety, therefore the scores are mainly below zero.

human readable name short explanation
Additive Scoring scoring_additive Sums up the scores of PSMs for the protein score. Usable when scores (instead of properties/e-values) are used, like Mascot ion score, SEQUEST XCorr
Multiplicative Scoring scoring_multiplicative Multiplies the scores of PSMs for the protein score. Usable when properties/e-values (instead of scores) are used, like Mascot or X!Tandem expect value orthe FDR Score
Geometric Mean Scoring geometric_mean_scoring Calculates the geometric mean of the PSMs' scores for the protein score. Usable when properties/e-values (instead of scores) are used.