SpectroPipeR - step 4 - statistics

If you want to have a closer look at the images and plots just do a “right-click” on the image and select “open in new tab”

The output folders contained figures that were always exported in two file formats: vector PDF and raster PNG.

Recommendations for MacOS Users

For MacOS users, it is recommended to use the PDF files, for example in Keynote, to achieve better quality.

Recommendations for Windows Users

For Windows users, it is recommended to use the PNG files, as applications like Microsoft PowerPoint can sometimes have issues when working with vector graphics.

SpectroPipeR statistics

The statistical analysis is carried out by utilizing the PECA package.

The recommended test is the “rots” test (ROPECA approach). This ROPECA (reproducibility-optimized peptide change averaging) approach can be inspected in detail in Enhanced differential expression statistics for data-independent acquisition proteomics.

background information

statistical analysis

Pair-wise comparison can be carried out using an ordinary t-test (“t”), modified t-test (“modt”), or reproducibility-optimized test statistic (“rots”). The type of data aggregation can be either median (“median”) or tukey (“tukey”) for calculating protein values. The test may be performed paired or unpaired depending on your experimental design.

PECA determines differential gene expression using directly the peptide intensity measurements of proteomic datasets. An change between two groups of samples is first calculated for each peptide in the datasets. The protein-level changes are then defined as median/tukey over the peptide level changes. For more details about the peptide-level expression change averaging (PECA) procedure, see Elo et al. (2005) doi:10.1093/nar/gni193, Laajala et al. (2009) doi:10.1186/gb-2009-10-7-r77 and Suomi et al. (2024) doi:10.18129/B9.bioc.PECA

PECA calculates the peptide level changes using the ordinary or modified t-statistic. The ordinary t-statistic is calculated using the function rowttests in the Bioconductor genefilter package.

If “modt” is selected modified t-statistic is calculated using limma package (fast processing time):

The modified t-statistic is calculated using the linear modeling approach in the Bioconductor limma package.

The empirical Bayes moderated t-statistics test each individual contrast equal to zero. For each protein, the moderated F-statistic tests whether all the contrasts are zero. The F-statistic is an overall test computed from the set of t-statistics for that peptide This is exactly analogous the relationship between t-tests and F-statistics in conventional anova, except that the residual mean squares have been moderated between proteins.

If “rots” is selected a Reproducibility-Optimized Test Statistic (ROTS) is calculated using the PECA package (slower processing time):

The reproducibility-optimization procedure (ROTS) enables the selection of a suitable gene ranking statistic directly from the given dataset. The statistic is optimized among a family of t-type statistics $$d_\alpha = |\overline{x}_1 - \overline{x}_2| / (\alpha_1 + \alpha_2*s)$$, where $|\overline{x}_1 - \overline{x}_2|$ is the difference between the two group averages of normalized peptide abundances, $\alpha_1$ and $\alpha_2$ are non-negative parameters to be optimized, and $s$ is the pooled standard error. The optimal statistic is determined by maximizing the reproducibility Z-score $$Z_k(d_\alpha) = (R_k * d_\alpha - R^0_k * d_\alpha) / s_k * d_\alpha$$ over a lattice of $\alpha_1\in\{0,0.01, ...,5\}$ and $\alpha_2\in\{0,1\}$, $k \in \{0,1,2,...,F\}$, where F is the total number of peptides in the data $R^0_k * d_\alpha$ is the corresponding reproducibility in randomized datasets permuted over samples and $s_k * d_\alpha$ is the standard deviation of the bootstrap distribution. Reproducibility is defined as the average overlap of $k$ top-ranked peptides over pairs of bootstrapped datasets. For protein-level inference of differential expression, the median of peptide-level p-values is used as a score for each protein taking the direction of change into account. The protein-level significance of the detection is then calculated using beta distribution. Under the null hypothesis, the p-values of the peptides follow the uniform distribution U(0,1). Furthermore, the order statistics from U(0,1) distribution follow a beta distribution. Finally, the FDR is calculated using the Benjamini-Hochberg procedure.

This complex and computation intense procedure allow a more precise estimate of significance than other methods. For further details please see Suomi & Elo 2017.

The significance of an expression change is determined based on the analytical p-value of the protein-level test statistic. Unadjusted p-values are reported along with the corresponding p-values looked up from beta ditribution. The quality control and filtering of the data (e.g. based on low intensity or peptide specificity) is left to the user.

Effect size implementation in SpectroPipeR

In SpectroPipeR the effect size (Cohen’s d) of peptide ratios per comparisons is implemented as follows. The peptide intensities were used to calculate the mean scaled peptide intensities per protein. This step is important since peptides have different intensities due to different flyability in mass spec. For example if two peptides are injected at 10fmol and one flies much better than the other, then the better flying peptide will have a much higher intensity. Therefore we need to adjust for this effect and scale the peptides in the same numeric region. Following the scaling process, the intensities of the scaled peptides were employed to compute Cohen’s d for each protein, facilitating subsequent comparative analysis.

example code

statistics_module() needs the output of the norm_quant_module() !

The condition_comparisons (mandatory) requires a cbind() with the user specified conditions as used during the setup of the analysis in Spectronaut. Therefore provide the conditions, which should be compared in the R code.

# condition comparison example
condition_comparisons_example <- cbind(
                                c("condition_1","condition_control"),
                                c("condition_2","condition_control"),
                                c("condition_3","condition_control")
                                )

# step 4: statistics module
SpectroPipeR_data_stats <- statistics_module(SpectroPipeR_data_quant = SpectroPipeR_data_quant,
                                       condition_comparisons = cbind(c("HYE mix A","HYE mix B")))

##*****************************************
## STATISTICS MODULE
##*****************************************
#
#reformatting data ...
#register processor cores ...
#performing statistical analysis (this might take a while) ...
#  |=============================================================================================| #100%
#close cores ...
#
#start to end time comparison for stat. analysis: 0.00111424499087863 hours
#estimating effect sizes ...
# [============================================================]  100.00% - calc. effect sizes... 
#
#join and tidy tables ...
#filtering statistical table using supplied cutoffs ...
#writing output files ...
#adding iBAQ quantiles to statistics table ...                                                                  # 
#generating Excel outputs ...                                                                                   # 
#performing fold-change cutoff sensitivity analysis ...                                                         # 
#plotting fold-change cutoff sensitivity analysis ...
# [================================================================================]  100.00% - 1 
#
#plotting fold-change simple cutoff sensitivity analysis (peptide n > 1)...
# [================================================================================]  100.00% - 1 
#
#generating volcano plots ...
# [============================================================]  100.00% - Volcano plots with adj. p-value 
# [============================================================]  100.00% - Volcano plots with adj. p-value 
# [============================================================]  100.00% - Volcano plots with raw p-value 
#
#condition-comparison-wise signal to noise comparison ...
#...signal to noise: save scatter plot...
#condition-comparison-wise comparison of peptide-int.-ratios vs. protein-int.-ratios ...
#...calculating protein ratios...
#...combining ratio tables...
#...generating ratio-ratios: protein_ratios/peptide_ratios...
#...filter for at least 2 peptides...
#...adding of protein intensity to table...
#...add signal to noise per group...
#...add detection with selected q-value cutoff with at least 2 peptides per replicate...
#...add direction comparison for protein or peptide condition comp. ratio...
#...save table of stat. significant with poor signal to noise...
#...counting protein which having a 2fold difference...
#...select Top15 over- or under-estimated proteins...
#...protein int. benchmark: save scatter plot...
#...protein int. benchmark: save histogram plot...
#...protein int. benchmark: save table...
#...counting protein: gradient of difference...                                                                 # 
#...protein int. benchmark: gradient of difference area plots...
#statistical analysis module done --> please check outputs in folder: ../SpectroPipeR_test_folder/06_statistics/  

statistics_module() outputs

The output in your specified output folder for the norm_quant_module() function should look like in this example (06_statistics, 05_processed_data):

statistics - figures

volcano plots

The volcano_plots_raw_p_value… illustrates the volcano plot (raw p-value) of the statistical analysis for a specific comparison.

The blue color indicates lower abundance and the orange indicates higher abundance of the protein in regard to their peptide ratios.

The blue, grey or orange label depicts the number of proteins in each fraction.

The user specified p-value and fold-change threshold are used to determine the fractions.

On the right panel of the plot the Top10 (based on euclidean distance) abundance differences are highlighted for the lower and upper fraction.

The volcano_plots_adjusted_p_value… illustrates the volcano plot (adjusted p-value / q-value) of the statistical analysis for a specific comparison.

The volcano_plots_effect_size_shape_adjusted_p_value… illustrates the volcano plot (adjusted p-value / q-value) of the statistical analysis for a specific comparison. The point shape depicts the estimated effect size.

cut-off plots

The cutoff_test… & cutoff_simple_test… illustrate the protein count by varying the fold-change threshold to filter significant proteins. They should help to estimate a meaningful project specific fold-change cutoff for the statistical analysis.

The simple cutoff plot highlights the protein count differences using various fold-change threshold with and without q-value filtering.

The cutoff plot highlights the protein count (with 1 or at least 2 peptides) differences using various fold-change threshold with and without q-value filtering

statistics - tables

statistical_analysis.csv

The statistical_analysis.csv holds the information of the statistical analysis.

• slr: signal log2-ratios on peptide basis
• t: t of t-statistics on peptide basis
• score: score of t-statistics on peptide basis
• n: number of peptides
• p: raw p-value of statistics on peptide basis
• p.fdr: adjusted p-value (q-value) of statistics on peptide basis
• PG.ProteinGroups: Protein groups
• group1: group1 of condition comparison
• group2: group2 of condition comparison
• slr_ratio_meta: condition comparison; how the ratio is formed
• test: which test was used for statistics on peptide level
• type: which type of ratio aggregation to ProteinGroup level was used for signal log2-ratios on peptide basis
• significant_changed: if there is a significant change FC & q-value (cutoffs e.g.: FC = 1.5 & adjusted-p-value = 0.05)
• significant_changed_raw_p: if there is a significant change FC & p-value (cutoffs e.g.: FC = 1.5 & p-value = 0.05)
• significant_changed_fc: fold-change cutoff used for analysis
• significant_changed_p_value: p-value/q-value cutoff used for analysis
• fold_change_absolute: ablsolute fold-change
• fold_change_direction: fold-change direction
• fold_change: fold-change
• effect_size_method: effect size estimation method used
• d: effect size estimate
• d_pooled_SD: effect size estimate; pooled SD
• d_95CI_lower: effect size estimate: the lower 95% confidence interval
• d_95CI_upper: effect size estimate: the upper 95% confidence interval
• d_magnitute: a qualitative assessment of the magnitude of effect size (|d|<0.2 negligible, |d|<0.5 small, |d|<0.8 medium, otherwise large); Cohen 1992

slr	t	score	n	p	p.fdr	PG.ProteinGroups	group1	group2	slr_ratio_meta	test	type	significant_changed	significant_changed_raw_p	significant_changed_fc	significant_changed_p_value	fold_change_absolute	fold_change_direction	fold_change	effect_size_method	d	d_pooled_SD	d_95CI_lower	d_95CI_upper	d_magnitute
-0.0473512	-0.4068343	0.6962885	2	0.7433455	0.9010067	A0PJW6	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.033366	down	-1.033366	Cohen’s d	-0.2791355	0.1074963	-1.3567385	0.7984675	small
0.0443521	0.4408390	0.6726308	9	0.8641528	1.0000000	A1X283	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.031220	up	1.031220	Cohen’s d	0.0164415	0.1860178	-0.4536598	0.4865428	negligible
-1.0186955	-21.5749594	0.0000001	1	0.0000001	0.0000003	A5Z2X5	B_manual	A_manual	B_manual/A_manual	modt	median	down	down	1.5	0.05	2.026086	down	-2.026086	Cohen’s d	-15.0652499	0.0450459	-24.4420988	-5.6884009	large
-0.1165372	-1.0692391	0.3204605	5	0.1911809	0.2713360	L0R6Q1	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.084130	down	-1.084130	Cohen’s d	-0.6380376	0.1449498	-1.2942932	0.0182179	medium
-0.0794428	-0.5016872	0.6312874	1	0.6312874	0.7874067	L0R8F8	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.056610	down	-1.056610	Cohen’s d	-0.3732878	0.1749949	-2.1185192	1.3719437	small
0.0066561	0.1008351	0.9225103	11	0.9999290	1.0000000	O00330	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.004624	up	1.004624	Cohen’s d	0.1792522	0.1897052	-0.2454271	0.6039315	negligible
-0.3342264	-2.2225784	0.0616644	1	0.0616644	0.0942844	O00458	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.260701	down	-1.260701	Cohen’s d	-1.5357510	0.1479843	-3.5045774	0.4330754	large
0.0079990	-0.0748142	0.9424568	10	0.9999709	1.0000000	O00487	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.005560	up	1.005560	Cohen’s d	0.0573662	0.3145541	-0.3878923	0.5026247	negligible
-0.0022888	-0.0321861	0.9752227	33	1.0000000	1.0000000	O00571	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.001588	down	-1.001588	Cohen’s d	-0.0370422	0.1975283	-0.2794375	0.2053530	negligible
-0.1181271	-0.6426256	0.5409473	9	0.5998730	0.7532240	O00622	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.085325	down	-1.085325	Cohen’s d	0.1723566	0.3216608	-0.2986087	0.6433219	negligible
-0.9757784	-8.2918171	0.0000727	9	0.0000000	0.0000000	O13516	B_manual	A_manual	B_manual/A_manual	modt	median	down	down	1.5	0.05	1.966702	down	-1.966702	Cohen’s d	-2.1848927	0.3573362	-2.7789090	-1.5908763	large
-1.2215582	-15.3916606	0.0000012	1	0.0000012	0.0000025	O13547	B_manual	A_manual	B_manual/A_manual	modt	median	down	down	1.5	0.05	2.331985	down	-2.331985	Cohen’s d	-7.9147819	0.1013292	-13.0563463	-2.7732176	large
-0.2271634	-2.2672522	0.0577325	1	0.0577325	0.0887239	O14521	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.170531	down	-1.170531	Cohen’s d	-1.5222502	0.0984701	-3.4871492	0.4426487	large
-0.0529473	-0.9991841	0.3510038	3	0.2831211	0.3925563	O14561	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.037382	down	-1.037382	Cohen’s d	-0.6291879	0.2031349	-1.4965383	0.2381626	medium
-0.0104878	-0.2441598	0.8141130	10	0.9892582	1.0000000	O14579	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.007296	down	-1.007296	Cohen’s d	0.4053804	0.1706905	-0.0443356	0.8550963	small

The files statistical_analysis_filtered… contain the same information but are filtered for:

file	filter criteria
statistical_analysis_filtered __fc_1.5__raw_p_value_0.05.csv	user-specified abs. fold-change (e.g. 1.5) and raw p-value (e.g. 0.05)
statistical_analysis_filtered __fc_1.5__raw_p_value_0.05 _2_more_peptides_per_protein.csv	user-specified abs. fold-change (e.g. 1.5) and raw p-value (e.g. 0.05) and at least 2 peptides
statistical_analysis_filtered __fc_1.5__adjusted_p_value_0.05.csv	user-specified abs. fold-change (e.g. 1.5) and adjusted p-value (e.g. 0.05)
statistical_analysis_filtered __fc_1.5__adjusted_p_value_0.05 _2_more_peptides_per_protein.csv	user-specified abs. fold-change (e.g. 1.5) and adjusted p-value (e.g. 0.05) and at least 2 peptides

statistical_analysis_2_more_peptides_per_protein_iBAQ_quantiles.xlsx

The Excel table statistical_analysis_2_more_peptides_per_protein_iBAQ_quantiles.xlsx (csv file also available) holds the information of the statistical analysis and the iBAQ quantiles associated with the comparisons.

iBAQ quantile ratio legend

• slr: signal log2-ratios on peptide basis
• iBAQ_quantile_comp: iBAQ quantiles of comparison
• t: t of t-statistics on peptide basis
• score: score of t-statistics on peptide basis
• n: number of peptides
• p: raw p-value of statistics on peptide basis
• p.fdr: adjusted p-value (q-value) of statistics on peptide basis
• PG.ProteinGroups: Protein groups
• group1: group1 of condition comparison
• group2: group2 of condition comparison
• slr_ratio_meta: condition comparison; how the ratio is formed
• test: which test was used for statistics on peptide level
• type: which type of ratio aggregation to ProteinGroup level was used for signal log2-ratios on peptide basis
• significant_changed: if there is a significant change FC & q-value (cutoffs e.g.: FC = 1.5 & adjusted-p-value = 0.05)
• significant_changed_raw_p: if there is a significant change FC & p-value (cutoffs e.g.: FC = 1.5 & p-value = 0.05)
• significant_changed_fc: fold-change cutoff used for analysis
• significant_changed_p_value: p-value/q-value cutoff used for analysis
• fold_change_absolute: ablsolute fold-change
• fold_change_direction: fold-change direction
• fold_change: fold-change
• effect_size_method: effect size estimation method used
• d: effect size estimate
• d_pooled_SD: effect size estimate; pooled SD
• d_95CI_lower: effect size estimate: the lower 95% confidence interval
• d_95CI_upper: effect size estimate: the upper 95% confidence interval
• d_magnitute: a qualitative assessment of the magnitude of effect size (|d|<0.2 negligible, |d|<0.5 small, |d|<0.8 medium, otherwise large); Cohen 1992
• group1__mean_iBAQ: mean iBAQ intensity of group 1
• group_1__iBAQ_quantiles: iBAQ intensity quantile of group 1
• group2__mean_iBAQ: mean iBAQ intensity of group 2
• group_2__iBAQ_quantiles: iBAQ intensity quantile of group 2

The statistical_analysis_2_more_peptides_per_protein_iBAQ_quantiles.xlsx holds information of the statistical analysis.

slr	iBAQ_quantile_comp	t	score	n	p	p.fdr	PG.ProteinGroups	group1	group2	slr_ratio_meta	test	type	significant_changed	significant_changed_raw_p	significant_changed_fc	significant_changed_p_value	fold_change_absolute	fold_change_direction	fold_change	effect_size_method	d	d_pooled_SD	d_95CI_lower	d_95CI_upper	d_magnitute	group1__mean_iBAQ	group_1__iBAQ_quantiles	group2__mean_iBAQ	group_2__iBAQ_quantiles
-0.0473512	Q4/Q5	-0.4068343	0.6962885	2	0.7433455	0.9010067	A0PJW6	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.033366	down	-1.033366	Cohen’s d	-0.2791355	0.1074963	-1.3567385	0.7984675	small	2588.925	Q4	2660.450	Q5
0.0443521	Q3/Q4	0.4408390	0.6726308	9	0.8641528	1.0000000	A1X283	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.031220	up	1.031220	Cohen’s d	0.0164415	0.1860178	-0.4536598	0.4865428	negligible	1404.825	Q3	1324.475	Q4
-0.1165372	Q8/Q9	-1.0692391	0.3204605	5	0.1911809	0.2713360	L0R6Q1	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.084130	down	-1.084130	Cohen’s d	-0.6380376	0.1449498	-1.2942932	0.0182179	medium	28498.150	Q8	31907.450	Q9
0.0066561	Q5/Q6	0.1008351	0.9225103	11	0.9999290	1.0000000	O00330	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.004624	up	1.004624	Cohen’s d	0.1792522	0.1897052	-0.2454271	0.6039315	negligible	4894.400	Q5	4884.950	Q6
0.0079990	Q8/Q9	-0.0748142	0.9424568	10	0.9999709	1.0000000	O00487	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.005560	up	1.005560	Cohen’s d	0.0573662	0.3145541	-0.3878923	0.5026247	negligible	30011.075	Q8	32368.550	Q9
-0.0022888	Q10/Q10	-0.0321861	0.9752227	33	1.0000000	1.0000000	O00571	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.001588	down	-1.001588	Cohen’s d	-0.0370422	0.1975283	-0.2794375	0.2053530	negligible	104792.700	Q10	112171.575	Q10
-0.1181271	Q5/Q5	-0.6426256	0.5409473	9	0.5998730	0.7532240	O00622	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.085325	down	-1.085325	Cohen’s d	0.1723566	0.3216608	-0.2986087	0.6433219	negligible	3684.925	Q5	3548.625	Q5
-0.9757784	Q10/Q10	-8.2918171	0.0000727	9	0.0000000	0.0000000	O13516	B_manual	A_manual	B_manual/A_manual	modt	median	down	down	1.5	0.05	1.966702	down	-1.966702	Cohen’s d	-2.1848927	0.3573362	-2.7789090	-1.5908763	large	138334.725	Q10	251255.500	Q10
-0.0529473	Q8/Q9	-0.9991841	0.3510038	3	0.2831211	0.3925563	O14561	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.037382	down	-1.037382	Cohen’s d	-0.6291879	0.2031349	-1.4965383	0.2381626	medium	26367.525	Q8	26587.425	Q9
-0.0104878	Q9/Q9	-0.2441598	0.8141130	10	0.9892582	1.0000000	O14579	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.007296	down	-1.007296	Cohen’s d	0.4053804	0.1706905	-0.0443356	0.8550963	small	44033.250	Q9	43786.775	Q9
-0.0820429	Q5/Q6	-1.0571006	0.3255960	6	0.1787867	0.2559203	O14657	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.058516	down	-1.058516	Cohen’s d	0.1920303	0.2575590	-0.3903803	0.7744409	negligible	4962.650	Q5	4795.525	Q6
-0.0063245	Q10/Q10	-0.0259550	0.9800181	14	1.0000000	1.0000000	O14818	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.004393	down	-1.004393	Cohen’s d	-0.0140356	0.2473297	-0.3885587	0.3604874	negligible	163847.650	Q10	162511.925	Q10
-0.0442411	Q3/Q4	-0.8248106	0.4366954	2	0.4196139	0.5551758	O14893	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.031141	down	-1.031141	Cohen’s d	-0.2370409	0.1437852	-1.3131936	0.8391119	small	1305.125	Q3	1411.100	Q4
0.0491310	Q8/Q9	0.5638555	0.5904612	12	0.7406270	0.8992711	O14929	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.034641	up	1.034641	Cohen’s d	0.6283915	0.1706648	0.2132161	1.0435668	medium	32077.275	Q8	30418.775	Q9
0.0534097	Q8/Q9	1.1650318	0.2821887	2	0.2317136	0.3251779	O14949	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.037715	up	1.037715	Cohen’s d	0.9962874	0.0374448	-0.1406891	2.1332639	large	26784.625	Q8	25997.875	Q9
-0.0265630	Q3/Q3	-0.2641864	0.7992497	12	0.9907917	1.0000000	O14981	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.018583	down	-1.018583	Cohen’s d	-0.2876621	0.2581572	-0.6950461	0.1197219	small	1097.700	Q3	1053.100	Q3

statistical_analysis_WIDE_FORMAT_2_more_peptides_per_protein.csv/xlsx

The statistical_analysis_WIDE_FORMAT_2_more_peptides_per_protein contains the information of the statistical analysis in wide-tabular format.

• “comparison”;signal_log2_ratio: signal log2-ratios on peptide basis of the specific comparison
• “comparison”;raw_p_value: raw p-value of the specific comparison
• “comparison”;adjusted_p_value: adjusted p-value of the specific comparison

PG.ProteinGroups	B_manual/A_manual|signal_log2_ratio	B_manual/A_manual|raw_p_value	B_manual/A_manual|adjusted_p_value
A0PJW6	-0.0473512	0.7433455	0.9010067
A1X283	0.0443521	0.8641528	1.0000000
L0R6Q1	-0.1165372	0.1911809	0.2713360
O00330	0.0066561	0.9999290	1.0000000
O00487	0.0079990	0.9999709	1.0000000
O00571	-0.0022888	1.0000000	1.0000000
O00622	-0.1181271	0.5998730	0.7532240
O13516	-0.9757784	0.0000000	0.0000000
O14561	-0.0529473	0.2831211	0.3925563
O14579	-0.0104878	0.9892582	1.0000000
O14657	-0.0820429	0.1787867	0.2559203
O14818	-0.0063245	1.0000000	1.0000000
O14893	-0.0442411	0.4196139	0.5551758
O14929	0.0491310	0.7406270	0.8992711
O14949	0.0534097	0.2317136	0.3251779

Protein intensity estimation comparison

Since bottom-up proteomics measures ions/peptides and not directly proteins the protein intensity estimation is slightly biased depending on the algorithm used. The ROPECA statistics used peptide ratios which can be compared to protein intensity ratios. This ratio comparison give researchers insights into the performance and agreement between various protein intensity estimation algorithms used in their proteomics data analysis. This information can be valuable for selecting the most appropriate estimation method and protein candidates per comparison.

Processed data - figures

Protein_intensity_benchmark__scatter_plot

The Protein_intensity_benchmark__scatter_plot plot depicts the peptide intensity ratio on the x-axis and the protein intensity ratio on the y-axis for proteins with 2 or more peptides.

The solid line shows the diagonal and the dashed lines indicate a 2-fold change. The count highlights the number of proteins which are affected by a difference above abs. 2 fold-change. The size of the dots illustrate the number of peptides. Usually proteins with low number of peptides are stronger affected by a difference between protein intensity estimation ratios, e.g. MaxLFQ ratios, vs. peptide ratios.

Protein_intensity_benchmark__MA_like_plot

The Protein_intensity_benchmark__MA_like_plot plot depicts an MA-like plot of the mean protein intensity (x-axis) and the peptide ratio / protein ratio on the y-axis. Usually proteins in the low abundant range are stronger affected by a difference between protein intensity estimation ratios and peptide ratios.

The solid line shows the diagonal and the dashed lines indicate a 2fold difference. The count highlights the number of proteins which are affected by a difference above abs. 2 fold-change. The size of the dots illustrate the number of peptides. Usually proteins with low number of peptides are stronger affected by a difference between protein intensity estimation ratios, e.g. MaxLFQ ratios, and peptide ratios.

Protein_intensity_benchmark__histogram_plot

The Protein_intensity_benchmark__histogram_plot plot depicts a histogram of the the peptide ratio / protein ratio on the x-axis.

Protein_intensity_benchmark__FC_gradient_area_plot

The Protein_intensity_benchmark__FC_gradient_area_plot presents an area plot across a range of fold-changes. In this instance, at a 1.5-fold change threshold, there are 76 proteins where the protein ratio is more than 1.5-fold to high, and 2 proteins where the protein ratio is more than 1.5-fold to low when compared to the peptide ratios.

Protein_intensity_benchmark__barplot

The Protein_intensity_benchmark__barplot displays a stacked bar plot, summarizing the number of proteins that exhibit a protein intensity ratio exceeding a 2-fold difference when compared to the peptide ratio.

processed data - tables

Protein_intensity_benchmark__table.csv

The Protein_intensity_benchmark__table.csv table holds all information for the protein intensity ratio vs. peptide intensity ratios on protein level.

annotation

parameter	description
slr	signal log2-ratios on peptide basis
t	t of t-statistics on peptide basis
score	score of t-statistics on peptide basis
n	number of peptides
p	raw p-value of statistics on peptide basis
p.fdr	adjusted p-value (q-value) of statistics on peptide basis
PG.ProteinGroups	Protein groups
group1	group1 of condition comparison
group2	group2 of condition comparison
slr_ratio_meta	how the ratio is formed
test	which test was used for statistics on peptide level
type	which type of ratio aggregation to ProteinGroup level was used for signal log2-ratios on peptide basis
significant_changed	if there is a significant change FC & q-value (cutoffs e.g.: FC = 1.5 & adjusted-p-value = 0.05)
significant_changed_raw_p	if there is a significant change FC & p-value(cutoffs e.g.: FC = 1.5 & p-value = 0.05)
significant_changed_fc	fold-change cutoff used for analysis
significant_changed_p_value	p-value/q-value cutoff used for analysis
fold_change_absolute	ablsolute fold-change
fold_change_direction	fold-change direction
fold_change	fold-change
median_protein_data_ratio	ratio of protein data (median protein intensity per condtion; ratio between 2 condtion medians)
log2_median_protein_data_ratio	log2 ratio of protein data (median protein intensity per condtion; ratio between 2 condtion medians)
log2_protein_ratios__peptide_ratios	log2 RATIOprotein/RATIOpeptide per comparison and ProteinGroup
protein_estimation_comment	is the protein intensity estimate ratio to high or to low in comparsion to the peptide ratio
group1_median_protein_intensity	median protein intensity estimate of group 1
group2_median_protein_intensity	median protein intensity estimate of group 2
group1_median_S2N	group1 median over signal-to-noise of ions
group2_median_S2N	group2 median over signal-to-noise of ions
group1_present_replicate_percentage	group1 present with selected q-value cutoff with at least 2 peptides per replicate: percentage of replicates per condition; NA = only 1 peptide per replicate >> may have globally 2 or more peptide
group2_present_replicate_percentage	group2 present with selected q-value cutoff with at least 2 peptides per replicate: percentage of replicates per condition; NA = only 1 peptide per replicate >> may have globally 2 or more peptide
direction_of_protein_ratios_and_peptide_ratios	does the protein ratio and peptide ratio point into the same direction?
effect_size_method	effect size estimation method used (mean scaled peptides intensities are used as input)
d	effect size estimate
d_pooled_SD	within-groups standard diviation
d_95CI_lower	the lower 95% confidence interval
d_95CI_upper	the upper 95% confidence interval
d_magnitute	a qualitative assessment of the magnitude of effect size (&#124;d&#124;<0.2 negligible, &#124;d&#124;<0.5 small, &#124;d&#124;<0.8 medium, otherwise large); Cohen 1992

table

slr	t	score	n	p	p.fdr	PG.ProteinGroups	group1	group2	slr_ratio_meta	test	type	significant_changed	significant_changed_raw_p	significant_changed_fc	significant_changed_p_value	fold_change_absolute	fold_change_direction	fold_change	effect_size_method	d	d_pooled_SD	d_95CI_lower	d_95CI_upper	d_magnitute	median_protein_data_ratio	log2_median_protein_data_ratio	log2_protein_ratios__peptide_ratios	protein_estimation_comment	group1_median_protein_intensity	group2_median_protein_intensity	group1_median_S2N	group2_median_S2N	group1_present_replicate_percentage	group2_present_replicate_percentage	direction_of_protein_ratios_and_peptide_ratios
-0.0473512	-0.4068343	0.6962885	2	0.7433455	0.9010067	A0PJW6	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.033366	down	-1.033366	Cohen’s d	-0.2791355	0.1074963	-1.3567385	0.7984675	small	0.8362415	-0.2580085	-0.2106574	OK	13769.343	16465.750	11.862334	12.819728	100	100	same direction - down
0.0443521	0.4408390	0.6726308	9	0.8641528	1.0000000	A1X283	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.031220	up	1.031220	Cohen’s d	0.0164415	0.1860178	-0.4536598	0.4865428	negligible	0.9013207	-0.1498876	-0.1942397	OK	5831.712	6470.186	4.011297	5.433860	100	100	contrary directions
-0.1165372	-1.0692391	0.3204605	5	0.1911809	0.2713360	L0R6Q1	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.084130	down	-1.084130	Cohen’s d	-0.6380376	0.1449498	-1.2942932	0.0182179	medium	0.7927020	-0.3351495	-0.2186123	OK	22810.444	28775.561	11.717550	8.904539	100	100	same direction - down
0.0066561	0.1008351	0.9225103	11	0.9999290	1.0000000	O00330	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.004624	up	1.004624	Cohen’s d	0.1792522	0.1897052	-0.2454271	0.6039315	negligible	0.8330499	-0.2635252	-0.2701814	OK	8532.935	10243.006	8.159923	9.367225	100	100	contrary directions
0.0079990	-0.0748142	0.9424568	10	0.9999709	1.0000000	O00487	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.005560	up	1.005560	Cohen’s d	0.0573662	0.3145541	-0.3878923	0.5026247	negligible	0.8300690	-0.2686968	-0.2766957	OK	21651.349	26083.793	9.446892	11.970736	100	100	contrary directions
-0.0022888	-0.0321861	0.9752227	33	1.0000000	1.0000000	O00571	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.001588	down	-1.001588	Cohen’s d	-0.0370422	0.1975283	-0.2794375	0.2053530	negligible	0.8616703	-0.2147921	-0.2125032	OK	53132.494	61662.205	12.812127	13.289387	100	100	same direction - down
-0.1181271	-0.6426256	0.5409473	9	0.5998730	0.7532240	O00622	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.085325	down	-1.085325	Cohen’s d	0.1723566	0.3216608	-0.2986087	0.6433219	negligible	0.8364571	-0.2576366	-0.1395095	OK	5735.594	6857.009	3.709370	4.977050	100	100	same direction - down
-0.9757784	-8.2918171	0.0000727	9	0.0000000	0.0000000	O13516	B_manual	A_manual	B_manual/A_manual	modt	median	down	down	1.5	0.05	1.966702	down	-1.966702	Cohen’s d	-2.1848927	0.3573362	-2.7789090	-1.5908763	large	0.4166547	-1.2630758	-0.2872973	OK	17678.528	42429.683	7.777781	14.886838	100	100	same direction - down
-0.0529473	-0.9991841	0.3510038	3	0.2831211	0.3925563	O14561	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.037382	down	-1.037382	Cohen’s d	-0.6291879	0.2031349	-1.4965383	0.2381626	medium	0.7971930	-0.3269991	-0.2740518	OK	15594.825	19562.170	10.766962	10.420646	100	100	same direction - down
-0.0104878	-0.2441598	0.8141130	10	0.9892582	1.0000000	O14579	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.007296	down	-1.007296	Cohen’s d	0.4053804	0.1706905	-0.0443356	0.8550963	small	0.8398016	-0.2518795	-0.2413917	OK	38715.729	46101.042	11.925857	13.447613	100	100	same direction - down
-0.0820429	-1.0571006	0.3255960	6	0.1787867	0.2559203	O14657	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.058516	down	-1.058516	Cohen’s d	0.1920303	0.2575590	-0.3903803	0.7744409	negligible	0.8233938	-0.2803455	-0.1983026	OK	10757.665	13065.031	4.936353	5.980577	100	100	same direction - down
-0.0063245	-0.0259550	0.9800181	14	1.0000000	1.0000000	O14818	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.004393	down	-1.004393	Cohen’s d	-0.0140356	0.2473297	-0.3885587	0.3604874	negligible	0.8519391	-0.2311779	-0.2248534	OK	40482.118	47517.622	17.675102	17.970990	100	100	same direction - down
-0.0442411	-0.8248106	0.4366954	2	0.4196139	0.5551758	O14893	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.031141	down	-1.031141	Cohen’s d	-0.2370409	0.1437852	-1.3131936	0.8391119	small	0.8145117	-0.2959927	-0.2517516	OK	8520.376	10460.717	6.269485	6.261318	100	100	same direction - down
0.0491310	0.5638555	0.5904612	12	0.7406270	0.8992711	O14929	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.034641	up	1.034641	Cohen’s d	0.6283915	0.1706648	0.2132161	1.0435668	medium	0.8979881	-0.1552317	-0.2043627	OK	19027.330	21188.843	10.134418	8.927515	100	100	contrary directions
0.0534097	1.1650318	0.2821887	2	0.2317136	0.3251779	O14949	B_manual	A_manual	B_manual/A_manual	modt	median	none	none	1.5	0.05	1.037715	up	1.037715	Cohen’s d	0.9962874	0.0374448	-0.1406891	2.1332639	large	0.8800108	-0.1844069	-0.2378166	OK	72653.469	82559.746	37.760534	50.907383	100	100	contrary directions

The Protein_intensity_benchmark__table_top15.csv table holds the information of the Top15 most deviating ratio / ratio comparisons.