Improvements in LC-MS/MS methods and technology have enabled the identification of thousands of modified peptides in a single proteomics experiment. However, to fully understand the role of such post-translational modifications (PTMs), robust quantification methods are crucial to identify and monitor the modification extent. Here, we introduce FLEXIQuant-LF, a software tool for large-scale identification of differentially modified peptides and quantification of their modification extent without prior knowledge of the type of modification. We developed FLEXIQuant-LF using label-free quantification of unmodified peptides and random sample consensus (RANSAC)-based robust linear regression to quantify the modification extent of peptides. The unbiased approach of FLEXIQuant-LF to assess the modification extent in quantitative proteomics data provides a novel platform to better understand the function and regulation of PTMs in new experiments and reanalyzed data.
Windows 10 and 270MB of free disk space.
The FLEXIQuant-LF GUI and FLEXIQuant-LF CLI executables as well as the Python source code can be downloaded here.
No installation needed.
FLEXIQuant-LF requires a comma separated file (.csv) containing peptide intensities in the following format as input:
| Column | Column header | Description |
|---|---|---|
| 1 | Sample | Name of sample (unique sample name) |
| 2 | Group | Name experimental group (will be used to identify reference samples) |
| 3 - n | * | Peptide intensities |
* can be anything, e.g. peptide sequence
Example file:
The example input file can be downloaded here.
FLEXIQuant-LF creates the following outputs:
-
_FQ-LF-output_raw_scores.csv: This file contains the calculated raw scores for all peptides.
-
_FQ-LF-output_RM_scores.csv: This file contains the calculated Relative Modification (RM) scores for all peptides that were removed by raw score filtering. RM scores are equivalent to one minus the modification extent.
Additionally, the following columns are included:
Slope: Slope of the regression line of the best linear regression model
R2 model: Coefficient of determination of the best linear regression model (calculated only using inliers as determined by RANSAC)
R2 data: Coefficient of determination of the best linear regression model (calculated using all data points) -
_FQ-LF-output_diff_modified.csv: This file states for each peptide if it was classified as differentially modified based on the chosen modification cutoff (True) or not (False)
-
_FQ-LF-output_removed_peptides.csv: This file lists the peptides that were remove by raw score filtering i.e. that resulted in a raw score higher than the median raw score of the sample plus three times the median absolute deviation of all raw scores of the sample
-
_FQ-LF-output_regression_plots.pdf (optionally): Contains a linear regression plot for each sample. Can be used as quality control the check if the linear regression line was fitted correctly.
-
Input file: Path to input file
-
Output folder: Path to output folder
-
Reference sample identifier: Based on this, FLEXIQuant-LF determines which sample(s) to use as reference sample(s). This needs to match exactly the value in the "Group" column for all samples that shall be taken as reference. If more than one sample is given (e.g. in case of a control group) FLEXIQuant-LF calculates the median intensity over all reference samples for each peptide.
-
RANSAC initiations: Number of times FLEXIQuant-LF fits a new RANSAC linear regression model to each sample to choose the best model. Between 5 and 100 initiations can be selected. The more initiations, the higher the reproducibility and the probability that FLEXIQuant-LF finds the optimal model. However, choosing a high number can significantly increases the run time and more than 50 initiations rarely provide additional benefit. Default: 30
-
Modification cutoff: RM score cutoff used to classify peptides as differentially modified. Default: 0.5
-
Create plots: If checked, a linear regression plot for each sample is created. All plots will be saved in one pdf file (_FQ-LF-output_regression_plots.pdf)
-
Select input file
-
Select output fold
Folder of the selected input file will be selected automatically as output folder. If a different location is preferred, select a different location.
-
Enter reference sample identifier
Needs to match group column value of reference samples
-
Select number of RANSAC initiations
The default setting is 30
-
Select modification cutoff
The default setting is 0.5
-
Check “Create plots” checkbox if you want FLEXIQuant-LF to create linear regression plots
-
Click “Run”
Full usage:
FLEXIQuant_LF_CLI.exe -i PATH -o PATH -r TEXT [-n INTEGER] [-mc FLOAT] [-p]
Usage example for provided test file:
FLEXIQuant_LF_CLI.exe -i Downloads\FLEXIQuantLF-master\src\test_file_CDC27.csv -o Downloads\FLEXIQuantLF-master\src\ -r "0h" -n 30 -mc 0.5 -p
(Input and output paths need to be adapted to the location of the test file on your computer.)
Note: FLEXIQuant_LF_CLI.exe needs to be started as described above from within a console or script. It can't be started by double-clicking the executable file.
Required arguments:
-i, --input_file PATH
Path of the comma separated input file (.csv)
-o, --output_folder PATH
Path to the folder in which the output files should be placed
-r, --reference TEXT
Based on this, FLEXIQuant-LF determines which sample(s) to use as reference sample(s). This needs to match exactly
the value in the "Group" column for all samples that shall be taken as reference. If more than one sample is given
(e.g. in case of a control group) FLEXIQuant-LF calculates the median intensity over all reference samples for each
peptide.
Optional arguments:
-n, --num_init INTEGER
Number of times FLEXIQuant-LF fits a new RANSAC linear regression model to each sample to choose the best model.
This should be an integer between 5 and 100. The more initiations, the higher the reproducibility and the probability
that FLEXIQuant-LF finds the optimal model. However, choosing a high number can significantly increases the run
time and more than 50 initiations rarely provide additional benefit. The default value is 30
-mc, --mod_cutoff FLOAT
RM score cutoff used to classify peptides as differentially modified. Should be a a float between 0 and 1.
The default value is 0.5
-p, --create_plots
If selected a linear regression plot for each sample is created. All plots will be saved in one pdf file
-h, --help
Show help message and exit