It is software for fragment analysis (MLPA, QF-PCR etc.) data processing.
FragalyseQt version 0.4+ supports assays with up to 8 different dyes used simultaneously, selective channel hiding, non-Latin run names, can correctly handle damaged files, exports peaks locations, areas, FWHM's, size and channel names in CSV for any FSA or HID files and exports in CSV internal analysis data for FSA or HID files generated by ABI 3500 and SeqStudio series equipment. Peak areas are calculated assuming they are Gaussian. For noisy runs denoising and baseline correction options is available. All the things are done without rewriting any data inside files FragalyseQt works with!
Currently, only FSA and HID files are supported. I've tried to add SCF files support, but it supports only 4 fluorescence channels and can't store information about size standard used for current assay. This renders SCF file format quite obsolete and unusable for modern assays and thus, this program won't support it.
But, anyway, if you wish to add support for any file format, you may help to implement it by sending patches.
Currently it is helpful for capillary electrophoresis troubleshooting, peak detection and sizing and, thus is applicable for quite a wast amount of different work.
Because FragalyseQt analyses raw capillary electrophoresis data, you may obtain peak areas different from true ones if your peaks are crowded (like in TP-PCR or at allelic ladders). In this case you MUST use baseline correction and denoising option.
FragalyseQt makes no changes in analysed files, so if you wish to save analysis results - you must export them in CSV.
Currently, I plan to add bining support to allow automated genotyping using FragalyseQt.
You may either provide run files from machines I haven't tested FragalyseQt with, provide patches or support me financially. Either case, contact me by email ( dorif11 AT gmail DOT com ), Telegram ( t.me/nurgle ) or here, at GitHub. Every help is important!
GNU GPL v3 for versions prior to 0.4.x. GNU AGPL v3 for versions from 0.4 and up.
Install Python 3.8+, git and PyQt5/PyQt6/PySide6 (only these Python bindings for Qt are supported) in any way that is comfortable to you. After this, clone FragalyseQt from Git repository and install remaining requirements. I propose you to do this using pip:
git clone https://github.com/Dorif/fragalyseqt && cd fragalyseqt && pip install -r requirements.txt && python3 main.pyIf you get externally-managed-environment error, try:
pip install --break-system-packages -r requirements.txt && python3 main.pyIt actually won't break anything in normal case.
Alternatively, you may try to set up virtual environment or use pipx, but previously mentioned option is recommended.
If you are using something RHEL-based - read next chapters and may the luck be with you.
Prepare enviroment to run FragalyseQt (GCC and python3-dev are needed to build BioPython):
su -c "apt-get install -y git gcc python3-module-pip python3-module-scipy python3-module-numpy-testing python3-dev"Download and run FragalyseQt:
git clone https://github.com/Dorif/fragalyseqt && cd fragalyseqt && pip install -r requirements.txt && python3 main.pyPrepare enviroment to run FragalyseQt:
sudo yum install python-qt5 gitDownload and run FragalyseQt:
git clone https://github.com/Dorif/fragalyseqt && cd fragalyseqt && pip install -r requirements.txt && python3 main.pyThese systems have no PyQt5 in their repos, nor you can install PyQt5 at these systems freshly installed. To install PyQt5 you should first update pip and clean pip cache:
sudo pip3 install --upgrade pip && rm -rf .pip/cacheAfter that install packages needed for FragalyseQt to work:
pip3 install pyqt5 && sudo yum install gitDownload and run FragalyseQt:
git clone https://github.com/Dorif/fragalyseqt && cd fragalyseqt && pip install -r requirements.txt && python3 main.pyFirst, install Git and libraries needed to run FragalyseQt (assuming you have Python 3.9):
pkg install git graphics/py-pyqtgraph biology/py-biopython textproc/py-charset-normalizer py39-pip &&\\
pip-3.9 install pybaselinesNow, download FragalyseQt and run it:
git clone https://github.com/Dorif/fragalyseqt && cd fragalyseqt && python3.9 main.pyOr, if you wish to install all python libraries from PIP (but make sure they are not installed as native packages):
pkg install git py39-pip && git clone https://github.com/Dorif/fragalyseqt && cd fragalyseqt &&\\
pip-3.9 install -r requirements.txt && python3.9 main.pyFirst, install Git and libraries needed to run FragalyseQt:
pkg_add git py3-pip py3-qt5 py3-pyqtgraph py3-biopython py3-charset-normalizer py3-scipy && pip3 install pybaselinesNow, download FragalyseQt and run it:
git clone https://github.com/Dorif/fragalyseqt && cd fragalyseqt && python3 main.pyOr, if you wish to install all python libraries from PIP (but make sure they are not installed as native packages):
pkg_add git py3-pip && git clone https://github.com/Dorif/fragalyseqt && cd fragalyseqt &&\\
pip3 install -r requirements.txt && python3 main.pyFirst, install Git and libraries needed to run FragalyseQt. In this case it is essential to install pyqtgraph by pip, because native NetBSD package seems to be broken in some way (assuming you have quite recent version of NetBSD and python 3.11):
pkg_add git py311-pip py311-qt5 py311-biopython py311-charset-normalizer py311-scipy &&\\
pip3.11 install pyqtgraph pybaselinesNow, download FragalyseQt and run it:
git clone https://github.com/Dorif/fragalyseqt && cd fragalyseqt && python3.11 main.pyOr, if you wish to install all python libraries from PIP (but make sure they are not installed as native packages):
pkg_add git py311-pip && git clone https://github.com/Dorif/fragalyseqt && cd fragalyseqt && \\
pip3.11 install -r requirements.txt && python3.11 main.pyYes, it will, because it is cross-platform not only in terms of OS, but also in terms of hardware. At least it was run at RISC-V and ARM based machines.
OpenVMS - has no Qt, no stable Python at least for x86_64 version, forget it. Solaris/Illumos - SciPy cannot be built and installed. Haiku - pyqtgraph package seems to be broken (resolves with installation by pip), SciPy package is compiled without ndimage module and fails to install by pip, maybe this will be fixed in future.
If you are using Linux or macOS, you may install Pyston (a performance-optimizing JIT for Python) for this purpose, it will speed up Python 3.7 - 3.10 compatible applications, not only FragalyseQt:
pip3 install pyston_lite_autoloadIt may be counterintuitive, but sometimes you MUST use size standard other, than detected and shown on screen right above CE data plot.
Here are some hints how to select correct size standard:
If you see LIZ/CC5/WEN/BTO/other channel 5 ladder is written above data plot, but no LIZ/CC5/WEN/BTO channel data are available - choose ROX/CXR/other channel 4 ladder upon your kit and equipment specifications. Example: all sample files for Profiler Plus, CoFiler and PowerPlex 16.
If you see limited ladder subset like GS600LIZ(60-460), but you see peaks obviously bigger than this ladder upper limit - count peak number and select ladder with highest peak size among counted. Example: in SeqStudio GlobalFiler files, shipped as example, GS600LIZ(60-460) ladder is detected, but in fact there are all ladder peaks up to 600 nucleotides, so you MUST use either GS600LIZ, GS600LIZ(60-600), GS600LIZ(80-600) or GS600LIZ(100-600) ladder for sizing.
Choose ladder subsets with skipped low size markers like GS600LIZ(60-600) if you have strong noise in that area, affecting even ladder channel.
Choose minus peak ladders subsets like GS600LIZ-250 ONLY if correspondent peaks failed to inject.
Fastest and least resource consuming with moderate precision is Linear Spline. Choose it if you have extremely weak system like an old netbook or first Raspberry Pi.
Most balanced is Cubic Spline.
5th degree spline may be a bit more precise than Cubic.
However, Cubic and 5th degree Splines are suffering from Runge phenomenon, making them about unusable for very small or very big sizes.
If you wish more sizing uniformity - use LSQ weighted spline sizing. Its precision grows from 1st to 5th order. The same way grows resource consumption.
Hit "SizeCall" button each time you change peak detection settings, apply baseline correction and denoising, hide/show channel or open new file.
FSA files from ABI 3100 and HID files from ABI 3500 are from NCBI OSIRIS test file set. NCBI OSIRIS is public domain software.
NCBI OSIRIS source is available at: https://github.com/ncbi/osiris
NCBI OSIRIS home page: https://www.ncbi.nlm.nih.gov/osiris/
Files from ABI 310 are from MLPAinter project, licensed under GNU GPL v3.
MLPAinter source code is available at: https://code.google.com/archive/p/mlpainter/
MLPAinter publication full text is available at: https://bmcbioinformatics.biomedcentral.com/articles/10.1186/1471-2105-11-67
NIST STR files are from NIST Forensic DNA Open Datadset v. 1.3.0 also available at: https://data.nist.gov/od/id/mds2-2157/pdr:v/1.3.0 Copyright notice added in correspondent directory according with Fair Use of Other NIST Data/Works policy: https://www.nist.gov/open/copyright-fair-use-and-licensing-statements-srd-data-software-and-technical-series-publications Only HID files from NIST dataset are kept, all other files are removed.
SeqStudio for GlobalFiler and RapidHIT ID v1.x files are obtained by myself and contain no other persons data.
COrDis Plus files for Nanophore-05 and SeqStudio were provided by Dmitrii Romaniuk, ISB, Moscow and contain no personal data.
Superyears Honor 1616 files were provided by Vadim Ponkratov, "Litolab", Berdsk, Russia and contain no personal data.