QuPE
QuPE Features
Motivation
After conducting proteome wet lab experiments, it may be noticed that the gathered data and information is somehow cluttered and distributed across several places: Details of the experimental setup are recorded in a laboratory book, mass spectrometry datasets are stored in some folder on a local hard disk, and lists of identified peptides and proteins are saved someplace else. To summarize, the relevant data belonging to one experiment are often not organised in a common fashion. In addition, techniques like stable isotope labelling yield quantitative information, implying the necessity of the application of further statistical tests and data mining approaches. We introduce our software QuPE as a repository and algorithmic framework to store and analyse mass spectrometry based quantitative proteome experiments.
Design and Implementation
QuPE is implemented in Java and based on the Spring framework, which provides a centralized, automated configuration and wiring of the individual parts of the system. Our J2EE-compliant application is deployable on any Java application server. Following the three tier architecture model, the system is separated into a data access, bussines logic, and presentation layer. The user interface of the software is implemented using the Echo2 web framework. With the help of an Ajax-based rendering engine, this approach combines the advantages of a web-based solution with those of a desktop application. On the one hand, the application is hosted as a service and provided over the internet. This eliminates the need for installation and simplifies software maintenance and support. On the other hand, the interactivity and responsiveness of a desktop application are retained. We used an in-house developed software tool to build the bridge between our object-oriented data model and a relational database management system. JavaO2DBI utilises Hibernate to create the data access layer of QuPE consisting of the appropriate POJO and DAO classes as well as the corresponding mapping files. Hibernate in combination with the Spring framework provides transaction management and encapsulates database access.
Job and Tools Concept
QuPE provides an easily extensible and configurable job concept. Using XML, jobs consisting of one or more tools can be defined, where input and output types provided by the implementation of a tool determine the data a job is executed with. Due to specific interfaces, tools can announce their need for an interactive configuration. The job and tool concept allows the integration of routines written in R, a programming language, specifically designed for mathematical and statistical purposes.
Features
- Webrowser-based application using Web 2.0 technologies
- Extensive capabilities to securely store and organise experiments and complete projects (fine-grained application-based security, GPMS)
- Import of mass spectrometry data: mzData, mzXML, the Mascot generic format
- Data model adapted to suggestions made by the HUPO proteomics standards initiative (PSI)
- Integrated Mascot search engine, import of search results, e.g. as DTASelect-filter
- Framework supporting analysis of quantitative proteomics data, including:
- Quantification of stable-isotope labelled samples (15N, 13C, SILAC)
- Significance tests, analysis of variance
- Principal component analysis
- Cluster analysis + validation: Hierarchical cluster analysis, K-means, Neuralgas
QuPE Access
Availability
QuPE is publicly available under the following url: http://qupe.cebitec.uni-bielefeld.de. For secured access to the application - this is highly recommended - please use the following address: https://qupe.cebitec.uni-bielefeld.de. If you like to test QuPE using an anonymous account, please go to the login page, and follow the "Try it out..." link from the menu on the right side.
In the framework of a collaboration, we can support you in terms of data processing and analysis, and offer you access to our compute infrastructure.We would be very pleased to collaborate with you in a common project. Do not worry any longer about software installations, maintenance or data backups! We look forward to hearing from you!
QuPE Algorithms & Resources
Peak intensity prediction in mass spectra using machine learning method
Mass spectrometry (MS) is an indispensable technique for the fast analysis of proteins and peptides in complex biological samples. One key problem with the quantitative mass spectrometric analysis of peptides and proteins, however, is the fact that the sensitivity of MS instruments is peptide-dependent, leading to an unclear relationship between the observed peak intensity and the peptide concentration in the sample. Various labeling techniques have been developed to circumvent this problem, but are very expensive and time-consuming. A reliable prediction of peptide-specific sensitivies could provide a peptide-specific correction factor, which would be valuable for label-free absolute quantitation.
This package of scripts contains tools that are organized in a pipeline. These have been used in the context of the phd project "Peak intensity prediction in mass spectra using machine learning methods" to extract peptide peaks from MALDI spectra, normalize intensities, generate features for machine learning, as well as train and test the prediction. You can view the abstract or downloaded the thesis from Bielefeld University. This package is provided as is, no guarantees are given for any of its functions. There is some documentation included with it which should allow a computer scientist to use or adapt the included scripts and software as seen fit. For further questions, feel free to contact the author.
QuPE Documentation
QuPE Documentation
There are several sources of information about the QuPE system available:
- The QuPE publication provides a brief overview of the system and the graphical user interface.
- A step-by-step explanation of important worksteps of QuPE, from the creation of a new experiment, to the upload of data, to the conduction of statistical analyses can be found on the following webpage: QuPE documentation (extern).
- Based on three real-world datasets, an application study was conducted that provides some guidance on the computational analysis of isotope-labeled mass spectrometry-based quantiative proteomics data.
- For software developers, there is also comprehensive documentation of the API of the QuPE system available.
- Information on algorithms and resources related to problems in mass spectrometry and protein identification can be found here.