Data Preprocessing

Data Preprocessing

Data preprocessing steps prepare imported data for downstream analysis by filtering, transforming, normalizing, imputing, simplifying, or removing problematic samples and features.

Filter Peptides

Filter Peptides: Existing Proteins

Removes all peptides whose protein ID is not found in the given protein dataframe.

Filter Peptides: Existing Samples

Removes all peptides whose sample is not found in the given protein dataframe.

Filter Peptides: PEP Threshold

Removes all peptides whose PEP value is greater than the selected threshold.

The PEP threshold indicates the probability that an identified peptide was assigned incorrectly. It is defined as: $$\text{PEP} = P(\text{false positive identification} \mid \text{given hit})$$

The smaller the PEP value, the more reliable the peptide identification.

Filter Proteins

Filter Proteins: Missing Samples

Filter Proteins: #Values / Group

Filter Proteins: Specific IDs

Filter Proteins: Keep n Most Significant

Filter PSMs

Filter PSMs: PEP Threshold

Filter PSMs: Existing Proteins

Filter PSMs: Existing Samples

Filter Samples

Filter Samples: Proteins / Sample

Filter Samples: Missing Proteins

Filter Samples: Sum of Intensities

Filter Metadata

Filter Metadata: Existing Samples

Outlier Detection

Outlier Detection: PCA

Outlier Detection: Local Outlier Factor

Outlier Detection: Isolation Forest

Transformation

Transformation: Log

Transformation: Inversion

Normalisation

Normalisation: Z-Score

Normalisation: Total Sum

Normalisation: Median

Normalisation: Width Adjustment

Normalisation: Reference Protein

Imputation

Imputation: Min per Dataset

Imputation: Min per Protein

Imputation: Min per Sample

Imputation: per Protein

Imputation: kNN

Imputation: Normal Dist. Sampling

Simplification

Group Replicates