AlphaPept: a modern and open framework for MS-based proteomics
Be sure to check out other packages of our ecosystem:
- alphatims: Fast access to TimsTOF data.
- alphamap: Peptide level MS data exploration.
- alphapeptdeep: Predicting properties from peptides.
- alphapeptstats: Downstream analysis of MS data
- alphaviz: Vizualization of MS data.
- Download the latest installer here, install and click the shortcut on the desktop. A browser window with the AlphaPept interface should open. In the case of Windows Firewall asking for network access for AlphaPept, please allow.
- In the
New Experiment, select a folder with raw files and FASTA files. - Specify additional settings such as modifications with
Settings. - Click
Startand run the analysis.
See also below for more detailed instructions.
| Feature | Implemented |
|---|---|
| Type | DDA |
| Filetypes | Bruker, Thermo |
| Quantification | LFQ |
| Isobaric labels | None |
| Platform | Windows |
Linux and macOS should, in principle, work but are not heavily tested and might require additional work to set up (see detailed instructions below). To read Thermo files, we use Mono, which can be used on Mac and Linux. For Bruker files, we can use Linux but not yet macOS.
We highly recommend the Anaconda or Miniconda Python distribution, which comes with a powerful package manager. See below for additional instructions for Linux and Mac as they require additional installation of Mono to use the RawFileReader.
AlphaPept can be used as an application as a whole or as a Python Package where individual modules are called. Depending on the use case, AlphaPept will need different requirements, and you might not want to install all of them.
Currently, we have the default requirements.txt, additional
requirements to run the GUI gui and packages used for developing
develop.
Therefore, you can install AlphaPept in multiple ways:
- The default
alphapept - With GUI-packages
alphapept[gui] - With pacakges for development
alphapept[develop](alphapept[develop,gui]) respectively
The requirements typically contain pinned versions and will be
automatically upgraded and tested with dependabot. This stable
version allows having a reproducible workflow. However, in order to
avoid conflicts with package versions that are too strict, the
requirements are not pinned when being installed. To use the strict
version use the -stable-flag, e.g. alphapept[stable].
For end-users that want to set up a processing environment in Python,
the "alphapept[stable,gui-stable]" is the batteries-included-version
that you want to use.
It is strongly recommended to install AlphaPept in its own
environment. 1. Open the console and create a new conda environment:
conda create --name alphapept python=3.8 2. Activate the environment:
conda activate alphapept 3. Install AlphaPept via pip:
pip install "alphapept[stable,gui-stable]". If you want to use
AlphaPept as a package without the GUI dependencies and without strict
version dependencies, use pip install alphapept.
If AlphaPept is installed correctly, you should be able to import AlphaPept as a package within the environment; see below.
- Install the build-essentials:
sudo apt-get install build-essential. - Install AlphaPept via pip:
pip install "alphapept[stable,gui-stable]". If you want to use AlphaPept as a package withouth the GUI dependencies and strict version dependencies usepip install alphapept. - Install libgomp.1 with
sudo apt-get install libgomp1.
- Copy-paste the Bruker library for feature finding to your /usr/lib
folder with
sudo cp alphapept/ext/bruker/FF/linux64/alphapeptlibtbb.so.2 /usr/lib/libtbb.so.2.
- Install Mono from mono-project website Mono Linux. NOTE, the installed mono version should be at least 6.10, which requires you to add the ppa to your trusted sources!
- Install pythonnet with
pip install pythonnet>=2.5.2
- Install AlphaPept via pip:
pip install "alphapept[stable,gui-stable]". If you want to use AlphaPept as a package withouth the GUI dependencies and strict version dependencies usepip install alphapept.
Only supported for preprocessed files.
- Install brew and pkg-config:
brew install pkg-config - Install Mono from mono-project website Mono Mac
- Register the Mono-Path to your system: For macOS Catalina, open the configuration of zsh via the terminal:
- Type in
cdto navigate to the home directory. - Type
nano ~/.zshrcto open the configuration of the terminal - Add the path to your mono installation:
export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig:/usr/lib/pkgconfig:/Library/Frameworks/Mono.framework/Versions/Current/lib/pkgconfig:$PKG_CONFIG_PATH. Make sure that the Path matches to your version (Here 6.12.0) - Save everything and execute
. ~/.zshrc
- Install pythonnet with
pip install pythonnet>=2.5.2
- Redirect to the folder of choice and clone the repository:
git clone https://github.com/MannLabs/alphapept.git - Navigate to the alphapept folder with
cd alphapeptand install the package withpip install .(default users) or withpip install -e .to enable developers mode. Note that you can use the different requirements here aswell (e.g.pip install ".[gui-stable]")
Some functionality of AlphaPept is GPU optimized that uses Nvidia’s CUDA. To enable this, additional packages need to be installed.
- Make sure to have a working CUDA
toolkit installation
that is compatible with CuPy. To check type
nvcc --versionin your terminal. - Install cupy. Make sure to install the cupy
version matching your CUDA toolkit (e.g.
pip install cupy-cuda110for CUDA toolkit 11.0.
To access Thermo files, we have integrated RawFileReader into AlphaPept. We rely on Mono for Linux/Mac systems.
To access Bruker files, we rely on the
timsdata-library. Currently, only Windows is supported. For feature finding, we use the Bruker Feature Finder, which can be found in theextfolder of this repository.
- For developing with the notebooks, install the nbdev package (see the development requirements)
- To facilitate navigating the notebooks, use jupyter notebook
extensions. They can be called from a running jupyter instance like
so:
http://localhost:8888/nbextensions. The extensionscollapsible headingsandtoc2are very beneficial.
To use AlphaPept as a stand-alone program for end-users, it can be installed on Windows machines via a one-click installer. Download the latest version here.
It is possible to run AlphaPept in a docker container. For this, we
provide two Dockerfiles: Dockerfile_thermo and Dockerfile_bruker,
depending on which filetypes you want to analyse. They are split because
of drastically different requirements.
To run, navigate to the AlphaPept repository and rename the dockerfile
you want to use, e.g. Dockerfile_thermo to Dockerfile.
- Build the image with:
docker build -t docker-alphapept:latest . - To run use
docker run -p 8505:8505 -v /Users/username/Desktop/docker:/home/alphapept/ docker-alphapept:latest alphapept gui(Note that -v maps a local folder for convient file transfer) - Access the AlphaPept GUI via
localhost:8505in your browser. - Note 1: The Thermo Dockerfile is built on a Jupyter image, so you can
also start a jupyter instance:
docker run -p 8888:8888 -v /Users/username/Desktop/docker:/home/jovyan/ docker-alphapept:latest jupyter notebook --allow-root
- The Thermo dockerfile was tested on an M1-Mac. Resources were set to 18GB RAM and 2 CPUs, 200 GB disk
- It was possible to build the Bruker dockerfile with the platform tag
--platform linux/amd64. However, it was very slow and the Bruker file is not recommended for an M1-Mac. Windows worked nicely.
The documentation is automatically built based on the jupyter notebooks (nbs/index.ipynb) and can be found here:
An overview of the performance of different versions can be found here. We re-run multiple tests on datasets for different versions so that users can assess what changes from version to version. Feel free to suggest a test set in case.
AlphaPept is meant to be a framework to implement and test new ideas quickly but also to serve as a performant processing pipeline. In principle, there are three use-cases:
- GUI: Use the graphical user interface to select settings and process files manually.
- CMD: Use the command-line interface to process files. Useful when building automatic pipelines.
- Python: Use python modules to build individual workflows. Useful when building customized pipelines and using Python as a scripting language or when implementing new ideas.
For the windows
installation,
simply click on the shortcut after installation. The windows
installation also installs the command-line tool so that you can call
alphapept via alphapept in the command line.
Once AlphaPept is correctly installed, you can use it like any other python module.
from alphapept.fasta import get_frag_dict, parse
from alphapept import constants
peptide = 'PEPT'
get_frag_dict(parse(peptide), constants.mass_dict){'b1': 98.06004032687,
'b2': 227.10263342687,
'b3': 324.15539728686997,
'y1': 120.06551965033,
'y2': 217.11828351033,
'y3': 346.16087661033}
If alphapept is installed an a conda or virtual environment, launch this environment first.
To launch the command line interface use: * alphapept
This allows us to select different modules. To start the GUI use: *
alphapept gui
To run a workflow, use: * alphapept workflow your_own_workflow.yaml
An example workflow is easily generated by running the GUI once and
saving the settings which can be modified on a per-project basis.
- Create a settings-file. This can be done by changing the
default_settings.yamlin the repository or using the GUI. - Run the analysis with the new settings file.
alphapept run new_settings.yaml
Within Python (i.e., Jupyter notebook) the following code would be required)
from alphapept.settings import load_settings
import alphapept.interface
settings = load_settings('new_settings.yaml')
r = alphapept.interface.run_complete_workflow(settings)
This also allows you to break the workflow down in indiviudal steps, e.g.:
settings = alphapept.interface.import_raw_data(settings)
settings = alphapept.interface.feature_finding(settings)
Within the notebooks, we try to cover most aspects of a proteomics workflow:
- Settings: General settings to define a workflow
- Chem: Chemistry related functions, e.g., for calculating isotope distributions
- Input / Output: Everything related to importing and exporting and the file formats used
- FASTA: Generating theoretical databases from FASTA files
- Feature Finding: How to extract MS1 features for quantification
- Search: Comparing theoretical databases to experimental spectra and getting Peptide-Spectrum-Matches (PSMs)
- Score: Scoring PSMs
- Recalibration: Recalibration of data based on identified peptides
- Quantification: Functions for quantification, e.g., LFQ
- Matching: Functions for Match-between-runs
- Constants: A collection of constants
- Interface: Code that generates the command-line-interface (CLI) and makes workflow steps callable
- Performance: Helper functions to speed up code with CPU / GPU
- Export: Helper functions to make exports compatbile to other Software tools
- Label: Code for support isobaric label search
- Display: Code related to displaying in the streamlit gui
- Additional code: Overview of additional code not covered by the notebooks
- How to contribute: Contribution guidelines
- AlphaPept workflow and files: Overview of the worfklow, files and column names
If you have a feature request or a bug report, please post it either as an idea in the discussions or as an issue on the GitHub issue tracker. Upvoting features in the discussions page will help to prioritize what to implement next. If you want to contribute, put a PR for it. You can find more guidelines for contributing and how to get started here. We will gladly guide you through the codebase and credit you accordingly. Additionally, you can check out the Projects page on GitHub. You can also contact us via [email protected].
If you like the project, consider starring it!
If you use this project in your research, please cite:
Strauss, M.T., Bludau, I., Zeng, WF. et al. AlphaPept: a modern and open framework for MS-based proteomics. Nat Commun 15, 2168 (2024). https://doi.org/10.1038/s41467-024-46485-4