Tool:FastQt : a clone of FastQC in C++/Qt

Question

[sacha]

Hi,

I recoded fastQC application from Java to Qt/C++ just for fun :D .
It should be faster, but I didn't make a real benchmark.

It was really easy to port the java code to Qt . In most of the time I just did a copy paste! Of course, I did some changes to keep C++/Qt pattern rule. Currently it can load only Fastq file... I will make Fastq.gz working as soon as possible.
And there are only 4 kind of analysis . I only make my favorite analysis right now. If you need one, feel free to ask me. Code source available here : https://github.com/labsquare/fastQt License GPL like FastQC.

[sacha]

Version 0.1 released !

I m pleased to present you the first release of FastQt version 0.1 . This version is pretty similar than FastQC and probably faster . It supports fastq, fastq.gz, fastq.bz2 and fastq.xz . Be free to tell us if you have bugs or any suggestion. The goal of the 0.1 version was to have same features than fastQC. The next goal in version 0.2 will focus on improvement and new feature like "multi analysis" in the same view and "Console mode". You can Download source code and compile it under Linux by following this link : https://github.com/labsquare/fastQt

A Binary for Linux x86_64 can be found here .
You just have do download it and run it as follow:

 chmod +x fastqt-0.1-linux-x86_64.AppImage
 ./fastqt-0.1-linux-x86_64.AppImage

Labsquare Team !

[WouterDeCoster]

Is there a size limit to the reads? I assume the tool isn't suited for "third-generation sequencing" data such as PacBio and Oxford Nanopore?

[sacha]

sequences size you mean ? I tested yesterday a fastq with long reads. It was working. Give me some data, I will try

[WouterDeCoster]

Yes, reads of multiple kb. And note that those reads will have very different sizes in the same dataset. Some human WGS nanopore sequencing data is available from https://github.com/nanopore-wgs-consortium/NA12878

[pmarijon]

Hi,

I am a contributor of fastqt and I work on long-reads every day, it's work very wheel and I have a feeling it's faster than fastqc.

Thank for your interest.

[WouterDeCoster]

Great, thanks for clarifying!

Answer 1

[Istvan Albert]

Why just copy the functionality of FastQC instead of improving on it.

One of the greatest flaws of FastQC is that one has to open every single file individually - incredibly tedious.

You should make a version of FastQC that works like it but improves on its two main and fundamental flaws. The serial file-by-file nehavior and that it does not show you results until it finishes running.

Instead of a single file the program should take a directoy full of Fastq files, and if so it should produce a smaller panel for each fastq file and shows them next to one another in a tile.

The second important feature would be to show data even before it is done, say every 5 seconds draws a quick histogram and shows what is going on so far.

[GenoMax]

One of the greatest flaws of FastQC is that one has to open every single file individually - incredibly tedious.

Only if one is using the GUI.

Multiqc written by Phil Ewels will take a directory full of FastQC (and other NGS data analysis reports) and produce a nice summary. In fact that would be a nice chapter to add to the handbook. I will plan to do that this week.

[Istvan Albert]

I did not hear of this tool yet, looks like a different tool that has to run fastqc in turn - not to mention it is has quite the overhead to set up.

I am just saying one shouldn't need the installation of another tool to process more than one file -- software improvements should help us reduce the mental overhead not add another tool to the pipeline.

[sacha]

You are right ! Ofcourse I plan to improve it. For the beginning, I was just trying to make it work like FastQC. I will change the configuration to open multi file at the same time.

Answer 2

[sacha]

FastQt 0.2 has been released !

• New UI
• Drag & Drop
• Windows Support and binary avaible
• Command Line interface
• Exporting results as PNG / SVG
• AppImage : Download it and run it from every Linux. Compilation is not necessary

For Linux :

wget https://github.com/labsquare/fastQt/releases/download/0.2/fastqt-0.2-linux-x86_64.appimage
chmod +x fastqt-0.2-linux-x86_64.appimage
./fastqt-0.2-linux-x86_64.appimage

Package for ArchLinux is avaible here:
https://aur.archlinux.org/packages/fastqt/

For Windows :

installer:
https://github.com/labsquare/fastQt/releases/download/0.2/fastqt-0.2-win32-mingw32.exe

Portable version :
https://github.com/labsquare/fastQt/releases/download/0.2/fastqt-0.2-win32-mingw32-portable.zip

Feedback

Please report any bugs on github page : https://github.com/labsquare/fastQt/issues.
You can also help us! We need an icon to replace the stupid bird icon :D

Answer 3

[sacha]

If you have any suggestion for new feature and improvement, let me know here !

[GenoMax]

If paired-end data exists then make it so that even if one file is selected the other one is analyzed automatically.

[Istvan Albert]

not just analyzed but reported together

[GenoMax]

I just came back to add that but Istvan Albert beat me to it :)

Display the resulting graphs side by side as well (if possible). It is valuable to check them simultaneously.

[GenoMax]

And use something other than a red "X" to highlight tests worth a second glance.

[WouterDeCoster]

It's interesting to look at the quality features of a sample, but even more interesting to find differences in profile/quality between samples. If I'm analyzing a cohort, I would definitely like it to be flagged if some/a few samples behave different compared to the rest.

Answer 4

[sacha]

What about something like this ? A Mdi Area where you can arrange window as you want .

[sacha]

it's done ! Watch the following !

[Istvan Albert]

yes, that is pretty cool, but by default I would line them like tiles, left to right and then let the people drag the windows around if they want to.