Two-Dimensional Polyacrylamide Gel Electrophoresis (2D-PAGE) provides two-dimensional maps where proteins appear separated according to their isoelectric point (pI) and molecular weight (MW). Usually these maps are very complex (i.e., hundreds or thousands of spots can be present in each map), and characterized by a low reproducibility, which hinders the possibility to identify reliable biomarkers unless robust methods are applied. The analysis of different sets of 2D-PAGE maps (e.g., control vs. pathological or control vs. drug-treated samples) to identify candidate biomarkers (proteins under- or over-expressed in different conditions) is usually carried out through image analysis systems providing a so-called spot volume dataset where each sample corresponds to a map described by the optical densities of all the detected spots. The identification of candidate biomarkers can be therefore accomplished by comparing different maps by classical monovariate statistical tests applied spotwise, or by multivariate chemometric tools applied to the entire set of spots present on each map. Here, the most exploited multivariate techniques will be considered, ranging from pattern recognition to classification methods.
Chemometric Tools for 2D-PAGE Data Analysis
Robotti E.
Primo
;Cala Elisa;Marengo E.
2021-01-01
Abstract
Two-Dimensional Polyacrylamide Gel Electrophoresis (2D-PAGE) provides two-dimensional maps where proteins appear separated according to their isoelectric point (pI) and molecular weight (MW). Usually these maps are very complex (i.e., hundreds or thousands of spots can be present in each map), and characterized by a low reproducibility, which hinders the possibility to identify reliable biomarkers unless robust methods are applied. The analysis of different sets of 2D-PAGE maps (e.g., control vs. pathological or control vs. drug-treated samples) to identify candidate biomarkers (proteins under- or over-expressed in different conditions) is usually carried out through image analysis systems providing a so-called spot volume dataset where each sample corresponds to a map described by the optical densities of all the detected spots. The identification of candidate biomarkers can be therefore accomplished by comparing different maps by classical monovariate statistical tests applied spotwise, or by multivariate chemometric tools applied to the entire set of spots present on each map. Here, the most exploited multivariate techniques will be considered, ranging from pattern recognition to classification methods.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
