Pathomx is a workflow-based tool for the analysis of data. It is interactive, visual, extensible and free for any use.
This release brings:
To report bugs, or request features see our issue tracker on Github, to get tips see the online documentation, or for more information see the Pathomx project website.
The key to analysis with Pathomx is the workflow. In data analysis it is common to perform a number of repeated steps in order, taking outputs of one action and feeding them into the next step in the analysis. In many cases the conversion from the output of one application to the input of another must be performed manually. For complex datasets this can significantly extend the time taken from getting data to finding results. Pathomx aims to solve this problem through the use of reusable, customisable, workflows. Workflows allow a number of processing steps to be connected together into a complex, parallel, system.
To get started, simply click on the sidebar and select 'Import'. Here you can select the type of data to import using one of the provided import plugins. There are a number of example datafiles available on the Pathomx website.
Once you've imported a dataset, you can then proceed to analyse it. Click on the Analysis tab and select PCA to perform Principal Components Analysis. The Scores plot will be shown immediately, with the principal component weight plots available on the tabs below.
If you click back to the Home tab, you will now see the Workspace view - showing your data source connected to the PCA analysis app. This is how all analysis is performed in Pathomx, by connecting apps to data sources (or the outputs of other apps). You'll notice that these two have been connected automatically for you by making a best guess of what you want. If something is connected wrongly it is simple to fix - simply drag the yellow output port onto the orange input port.
If your input data is not mean centred you may have noticed the PCA weights plots are a not centred around a zero-point. To fix this we need to mean center the data before applying the PCA. Click on the Processing tab and select 'Transform' to launch a simple app providing basic data-transformation options. On the Transform toolbar select 'Mean centre' from the drop down menu (containing 'log2').
If you go back to your workspace you'll see that the application has guessed wrongly and connected the mean centring to the output of the PCA. It has made this assumption on the basis of the order you have added the applications. So correct the order by first dragging from the output of your data input to the input of the Mean Centre app, then from the output of the Mean Centre app to the input for PCA.
If you now return to the PCA app (click on it) you'll see the corrected, mean centered PCA.
While this has taken a few minutes to set up, you can save this workflow for future use by using File > Save Workflow As....
Another useful features is the inline views that are available from most tools. Simply right-click and select a view from the drop down list under Views, and it will be pinned onto the current workflow. These views automatically update as processing occurs, so you can get a instant overview of what is happening under the hood.
More information and demo workflows are available on the Pathomx website. We've also put together a number of demo videos to demonstrate these functions in more detail.
Pathomx is available free for any use, licensed under the GPLv3.
Pathomx is built on the MetaCyc pathway database itself part of the BioCyc and HumanCyc family. The supplied database is generated via the MetaCyc API and stored locally. Licenses for the entire MetaCyc database are also available free of charge for academic and government use.
Molecular structures are derived from the KEGG database, developed by Kanehisa Laboratories. If using molecular structure annotations please cite the following article(s): Kanehisa, M. and Goto, S.; KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res. 28, 27-30 (2000).
tbc.
Pathomx is under continued development. If you spot bugs, or have suggestions for improvement, please submit them via our github bug tracker.
Releases of Pathomx are available for download from our website packaged for both MacOS X and Windows x64 (XP, 7 and 8). Unpackaged source files are also available for manual installation or development from our github repository.
2.1.0 | Renamed 'Pathomx' due to name clash with existing software. Adds interactive configuration panels to tool windows. (04/02/2014) |
2.0.0 | New Visual workflow editor, Matplotlib-based graphing, improved interface, API improvements (12/01/2014) |
1.0.0 | Interactive workflow analysis support, plugins, Windows and Mac support, multithreading. (18/12/2013) |
0.7.0 | GPML visualisation support, data heatmaps introduced. (01/07/2013) |
0.6.0 | GPML support, compartmentalisation of reactions and molecular structure images introduced. (01/06/2013) |
0.5.2 | Fixing packaging via PyPi. Some small bugfixes for map generation speedups. (01/02/2013) |
0.5.1 | Initial release supporting metabolic pathway exploration and annotation with metabolic data. Basic implementaiton of pathway pruning for identification of key metabolic networks within a dataset. (01/01/2013) |
Please direct all correspondence to Martin Fitzpatrick, Rheumatology Research Group, Centre for Translational Inflammation Research, Queen Elizabeth Hospital, Mindelsohn Way, Birmingham, B15 2WB, United Kingdom