Introduction to ProtE

Function inputs

ProtE features 4 functions, each one tailored for a specific use case.

1. maximum_quantum() accepts as input the MaxQuant generated file ProteinGroups.txt
2. dianno() accepts as input either of the two DIA-NN (or the FragPipe - DIANN) output files pg_matrix.tsv or unique_genes_matrix.tsv
3. pd_single() accepts as input the Proteome Discoverer output file that contains all sample protein intensities/abundances in one table
4. pd_multi() accepts as input separate Proteome Discoverer protein intensity files

How to use functions maximum_quantum(),dianno(),pd_multi()

All 3 functions expect the input file to be parsed in the parameter file. To enable statistical analysis, in the input file, samples (columns) belonging to the same group must be sorted next to each other. For example, samples from an experiment with a 3-groups categorical variable (control, treatment, compound) could be arranged such that: first columns = Control samples, middle columns = Treatment samples, last columns = Compound samples.

Seting up the input file path:

Assuming a MaxQuant quantification has been performed, the file ProteinGroups.txt can be fed to ProtE with the function maximum_quantum.

Insert the file path of the ProteinGroups.txt in the file parameter. To copy-paste the file path in Windows, firstly locate the desired file inside your folders. Hold Shift and right-click the file, then select “Copy as Path” from the context menu. Go to RStudio and click Ctrl+V or right-click to paste the path.

Because usually the directories will be separated with a single backlash, ensure to use forward slashes (/) for specifying paths or adding a second backlash e.g:

maximum_quantum(file = "C:\\Bioprojects\\BreastCancer\\Proteomics\\MaxQuant\\ProteinGroups.txt")

or

maximum_quantum(file = "C:/Bioprojects/BreastCancer/Proteomics/MaxQuant/ProteinGroups.txt")

Setting up group_names and number of samples_per_group

Group names are defined in the parameter group_names as a vector. The order of the group names inside the vector must follow the order of the groups by which the samples (columns) have been arranged in the input proteomics file (from the left to the right). Same goes for the number of samples of each group, which is defined again as a vector in the parameter samples_per_group. In the following example there are 3 groups (Control,Treatment,Compound) with the Control group consisting of 10 samples the Treatment group of 12 samples and the Compound with 9:

maximum_quantum(
                    file = "C:\\Bioprojects\\BreastCancer\\Proteomics\\MaxQuant\\ProteinGroups.txt",
                    group_names = c("Control", "Treatment", "Compound"),
                    samples_per_group = c(10, 12, 9),
                    imputation = FALSE,
                    global_filtering = TRUE,
                    independent = TRUE,
                    filtering_value = 50,
                    normalization = FALSE,
                    parametric= FALSE,
                    significance = "p")

In the pairwise comparisons, nominators and denominators of the FoldChange (and consequently the sign of Log2FoldChage) are defined based on the order of the group names declared in the parameter group_names. The general notion based on which FoldChange is determined is: NextGroup/PreviousGroup. In our example the FoldChange for every pairwise comparison will be set as: Treatment/Control, Compound/Control and Compound/Treatment.

How to use pd_multi()

pd_multi is tailored for the analysis of multiple Proteome Discoverer (PD) exports, each one corresponding to a single sample. To be able to use it, the user must save the PD exports to different folders corresponding to the different groups of the variable that is going to be analyzed. The paths to these folders are specified in the parameter …:

pd_multi(excel_file = "C:\\Bioprojects\\BreastCancer\\Proteomics\\PD\\Control",
                       "C:\\Bioprojects\\BreastCancer\\Proteomics\\PD\\Treatment",
                       "C:\\Bioprojects\\BreastCancer\\Proteomics\\PD\\Compound",
                    imputation = FALSE,
                    global_filtering = TRUE,
                    independent = TRUE,
                    filtering_value = 50,
                    normalization = FALSE,
                    parametric= FALSE,
                    significance = "p")

In the pairwise comparisons, nominators and denominators of the FoldChange (and consequently the sign of Log2FoldChage) are defined based on the order of the declared group folders in the pd_multi function. Again, the general notion based on which FoldChange is determined, is: NextGroup/PreviousGroup. In our imaginary example the FoldChange for every pairwise comparison will be set as: Treatment/Control, Compound/Control and Compound/Treatment.

Summary of the ProtE pipeline

All 4 functions streamline the following process, which is reported in more details in the ProtE Guide vignette.

1. Normalization of proteomic intensity values

2. Filtering based on the percentage of missing values of each protein.

3. Imputation of missing data to ensure robust downstream analysis.

4. Description fetching for DIA-NN’s .pg_matrix.tsv and Proteome Discoverer files

Once the data processing is complete, the package performs statistical analysis for every pairwise comparison to identify significant protein abundance differences between experimental groups. The results are automatically exported as Excel files, and a range of visualizations is generated to facilitate quality check and interpretation. These include:

• Principal Component Analysis (PCA) plots for dimensionality reduction and group comparison.

• Heatmap highlighting significant proteins.

• Protein rank-abundance and meanrank-sd scatterplots.

• Boxplots and violin plots to display data distribution and variability across groups.

Output Directory

The results from each function are saved in a folder named ProtE_Analysis, which is created inside the last directory of the provided file(s).

ProtE creates 3 sub-folders: • Data_processing, with the files of the resulting data processing. • Statistical_Analysis, with the results of the statistical tests. • Plots, with all the plots saved in bmp. format.