Table 2.

Batch effect processing checklist.

Step	Substeps
Experimental designa	Randomize samples in a balanced manner to prevent confounding of biological factors with batches (technical factors).
	Consider adding replicates if possible, for example: (a) add replication for each technical factor; (b) regularly inject a sample mix every few (e.g., 10–15, but the exact number will need to be adjusted depending on experimental conditions) samples for control; (c) incorporate a sample mix per batch.
	Record all technical factors, both plannable and occurring unexpectedly.
Initial assessment	Check whether the sample intensity distributions are consistent.
	Check the correlation of all sample pairs.
	If intensities or sample correlations differ, check whether the intensities show batch‐specific biases.
Normalization	Choose a normalization procedure, appropriate for biological background and data properties.
Diagnostics	Using diagnostic tools, determine whether batch effects persist in the data.
	Use quality control already at this step and skip the correction if it is not necessary.
	Tip: If the goal is to determine differentially expressed proteins, and the batch effects are discrete or linear, multi‐factor ANOVA on normalized data is a sound statistical approach. This will adjust for batch effects while simultaneously identifying differentially expressed proteins. Note, that "hits" or differentially expressed proteins identified with this approach are valid even if diagnostic tools indicate the presence of batch effects. For more details on ANOVA methods, refer to (Rice, 2006).
Batch effect correction	Choose batch effect correction procedure, appropriate for the biological background and data properties, especially those detected at the previous step.
	Repeat the diagnostic step.
	Assess the ultimate benefit with quality control.
Quality control	Compare correlation of samples within and between the batches. Pay special attention to replicate correlation, if these are available.
	Compare correlation of peptides within and between the proteins.

^a For details on experimental design, see (Čuklina et al, 2020).