- Open Access
Brunn: An open source laboratory information system for microplates with a graphical plate layout design process
© Alvarsson et al; licensee BioMed Central Ltd. 2011
- Received: 27 October 2010
- Accepted: 20 May 2011
- Published: 20 May 2011
Compound profiling and drug screening generates large amounts of data and is generally based on microplate assays. Current information systems used for handling this are mainly commercial, closed source, expensive, and heavyweight and there is a need for a flexible lightweight open system for handling plate design, and validation and preparation of data.
A Bioclipse plugin consisting of a client part and a relational database was constructed. A multiple-step plate layout point-and-click interface was implemented inside Bioclipse. The system contains a data validation step, where outliers can be removed, and finally a plate report with all relevant calculated data, including dose-response curves.
Brunn is capable of handling the data from microplate assays. It can create dose-response curves and calculate IC50 values. Using a system of this sort facilitates work in the laboratory. Being able to reuse already constructed plates and plate layouts by starting out from an earlier step in the plate layout design process saves time and cuts down on error sources.
- Survival Index
- Microplate Assay
- Laboratory Information System
- Control Marker
- Dialog Window
The life sciences of today is an information-heavy field with extensive use of high-throughput methods that generate large amounts of data at an ever-increasing rate [1, 2]. As the quantity of data increases, proper methods for handling the data become important. Good data handling with standardization is important not only in the analysis step, but also for the data generation step . Modeling wet-laboratory experiments in a computerized information system is a non-trivial task and needs a flexible system capable of supporting the daily laborative work related to microplates, which are the main media used in high-throughput assay systems.
General systems for ultra high throughput screening need highly optimized and streamlined platforms , whereas lead optimization systems need to be more flexible and handle a more fluctuating range of plate formats and layouts . This study was devoted to the development of such a flexible laboratory information system (LIS) for microplates, termed "Brunn". Bioclipse  was choosen as a platform to base the LIS on. The system is designed to help in refining raw data and encourages re-use of layouts for the plates, Brunn thus focuses on providing a platform for fluctuating plate layouts; here we demonstrate it on a fluorometric microculture cytotoxicity assay (FMCA) in use at the university hospital of Uppsala, Sweden.
Microplates were introduced in the early 1960s . Since then they have become an integral part of daily laboratory work in most disciplines, including drug discovery and clinical routine testing. Microplates consist of an array of wells. The outer dimension of microplates is standardized, but the plates are available with different numbers of wells and volumes. Microplates can be handled both manually and by robots, and they facilitate sample handling and serial dilutions by transferring solutions across rows or columns in parallel. They can be manufactured with different optical properties, such as for fluorescent read-outs, colorimetric assays, and microscopy. In addition, the plastics can be given different properties, e.g. making a sample adhere, or not adhere, to the walls of the well. Moreover, wells can be shaped in different ways (most common are flat and V-shaped bottoms). Microplates are available in a variety of standardized densities, where 8 rows × 12 columns = 96 wells with a volume of 200-300 μL are the most suited to be handled manually. Much denser plates are also available, among which 384-well plates can be handled manually, though this is rather painstaking. Plates with higher densities than this are difficult to use; one reason for this is that the small volumes of liquid in each well evaporate quite fast.
Microplates allow large quantities of data to be generated fast. Most assays give one result per well, but it is also possible to perform time-resolved assays with one measurement per time-point per well. In high-throughput settings, thousands of measurements are generated by automatic plate handling robots, of which many types are available on the market. Collecting high-throughput data in a laboratory information system gives great benefits. For example, assays commonly exhibit a systematic variation across the geometry of the plate, which necessitates controls to be distributed in a statistically secured way over the plates. In order to handle this effectively a LIS for microplates is required.
Bioclipse inherits the top-of-the-line plugin structure of Eclipse, which provides an extensible environment where each Bioclipse installation can be tailored with the features needed by a certain user. Choosing Bioclipse as the platform for Brunn means that Brunn does not need to reinvent the proverbial wheel, but can use suitable parts of Bioclipse already available, and it allows for the easy integration of the microplate assay data with the other features of Bioclipse. The Bioclipse workbench consists of editors and views. Editors follow the load/save pattern and many instances of the same editor can be open at the same time, whereas views reside around the editors and shows different information; for example a table of properties for the selected resource.
Brunn ships as a set of plugins for Bioclipse. For example, a core plugin contains the business logic, and a user interface plugin contains the graphical visualisation parts.
Resources in Brunn
A PlateType defines the size of a plate.
A PlateLayout defines which wells are used for dilution series, blank and controls. It also defines all calculation functions.
A MasterPlate defines which substances with which concentrations are in which wells.
A Compound is a substance whose effects are of interest.
A Cell Line represents a cultured cell type that is grown outside the body. Can be bred for specific properties such as drug resistance
A Patient Cell represents tissue samples from a patient.
A Plate corresponds to a real life plate and is what result data is coupled to.
Dose-response curves can also automatically be generated and IC50-values calculated.
Auditing and Users
Brunn has two user levels, administrators and regular users. An administrator can do what all regular users can do, as well as create new users and undelete resources which previously have been deleted. Brunn keeps track of who created what resource and when. When a resource is selected in the Brunn Explorer, the Properties View will list creator and creation date for that resource, as shown in Figure 2. After a resource has been created in Brunn, for security reasons, it can actually not be deleted - only marked as deleted.
Defining a PlateType
It is possible to create Plates of any size by defining a PlateType with the desired number of rows and columns. There are no restrictions in the data model for the sizes of the plates, but as mentioned above currently not all parts of Brunn support sizes other than 96- and 384-well microplates.
Creating a PlateLayout
When working with microtiter plates it is common to use the same general layout and calculate the same values for the same wells for multiple plates. In order for the user to only have to input the information once, a reusable PlateLayout can be created. PlateLayouts are created from PlateTypes. The PlateLayout editor has a point-and-click interface for placing well markers symbolizing different groups of control, blank and treatment wells. When control markers are placed, functions for average and standard deviation for the wells with those markers are automatically created. The PlateLayout editor also contains a user interface for creating other user-defined plate-specific functions.
In order to compensate for drift within the plate, different control wells can be used when calculating things like survival index for various wells. This means that well-functions, which are containers for well-specific calculation formulæ, for example SI %, need to be created manually only once for each group of wells using the same control.
Well functions and Plate functions make it possible for the user to create a PlateLayout, tailored to calculate exactly what the user is intrested in; for example survival index or various statistical measures used during primary screening. (See Figure 2 for an example of a Well function.)
Placing substances - defining a MasterPlate
Plates are normally prepared and cryostored in batches where all plates in one batch have the same substance with the same concentration at the same location. In a subsequent step, they are seeded with samples. Creating such a batch of plates corresponds to creating a MasterPlate in Brunn.
In the MasterPlate editor, the substance markers placed in the PlateLayout editor are associated with a substance with a different concentration for each well. This is done by dragging-and-dropping a substance to a marker and creating a dilution series in a dialog window.
Adding samples - creating a Plate
The seeding of a physical plate with samples corresponds to creating a Plate in Brunn. The entity Plate corresponds to a microplate with a barcode, substances, and samples seeded on it. The dialog window for creating a Plate lets the user specify barcode, plate name, cells to seed and a MasterPlate to base the new Plate on. Once a plate has been created, data can be imported to it.
Data browsing and curation
Once raw data has been imported, the Plate editor can be used for viewing the results for one Plate. The Plate editor has multiple tabs where data can be viewed in different ways (see Figure 2).
By using the proper tabs of the Plate editor, wells can be marked as outliers, effectively removing them from further calculations. To help the user identify outliers the calculation functions can be viewed. For example, when a plate layout is created with markers indicating wells used for blank and control values, functions like coefficient of variance as percent (CV %) for all wells with those markers are automaticly created, or if doing primary screening the user can write a plate function for standard deviation of all screened substances and for example use it when creating well functions, like Z-scores .
The last tab in the Plate editor shows a graphical report containing dose-response curves and automatic calculations of IC50. The dose response curves are drawn with straight lines between the points and the IC50 values are calculated by identifying the two points closest to 50% and then doing a linear aproximation between them to get a value for IC50. This is the default behavior of Brunn because when working with cell assays the shape of the curves tend to vary a lot. However, if the assay gives output suitable for a sigmoidal curve fitting there is a user configurable preference that toggles IC50-calculation by fitting the Hill equation using non-linear least-squares regression with the org.apache.commons.math  library.
Accessing the Brunn database in Bioclipse
Cheat Sheets in Brunn
Shows how a PlateLayout with layout markers and calculation functions is created.
Creating a MasterPlate
Shows how a MasterPlate is created and how substances are created and added to the MasterPlate.
Creating a Plate and Importing Results
Shows how a Plate is created, data imported for it and how the results can be viewed.
Example use case: FMCA
While Brunn was developed to be general enough to handle virtually any type of microplate assay, it was validated on a microplate-based fluorometric microculture cytotoxicity assay (FMCA) , an assay type that is in wide use for measuring cytotoxic and cytostatic effects of chemical compounds in vitro. It is a cell viability assay, and the specific one used herein is in use at the division of Clinical Pharmacology of the Uppsala University Hospital.
Figure 3 shows an overview of the process. Cells are seeded onto 96- or 384-well microplates with dilution series of substances. By measuring fluorescence generated from cells with intact cell membranes, a survival index (SI %) is calculated for each concentration of each substance. In order to accomplish this, some wells on the plate must act as blanks and controls. Moreover, to compensate for systematic variations along the plate, different wells are used as controls for different parts of the plate.
There is a trade-off between testing as large a number of substances as possible on one plate with enough number of replicates, and testing at a sufficient number of different concentrations. Hence, the layout of the plates (i.e. which wells are used for dilution series, and which wells are used for more technical ends such as controls and blanks) is continually improved, and there are multiple layouts being used side by side.
Control markers and Brunn
Does not handle positional variability
Biases may be added by the controls
Control variability is not accounted for
Not clear how to handle outliers
Brunn uses built in support for control markers for catching systematic errors and at least partly provides solutions for dealing with these potential problems. In the FMCA example the position variability problem (1) has been approached in two ways. First, there has been trouble with different values in the outermost wells on the plates and hence they have not been used for measurements. Second, there are different control markers on the plate and the closest one is used in the calculations. To keep track of control variability (3) the CV % is calculated both for each control group individually and all controls together.
These CV values can be used as a hint for identifying outliers (4). Still the outlier problem is always present and Brunn does not provide the final solution for how to handle outliers. It is important for the user to have issues like this in mind when designing the platelayout. Brunn provides a framework where things can be set up and organized in a systematic way.
Comparison of Brunn with other open source LIS
There are two other open source LIS systems for microplate assays which deserve to be mentioned, namely SLIMS  and Screensaver . SLIMS is dedicated for chemical genetics and handles large-scale screening data. Screensaver also handles high-throughput screening facilities. However, SLIMS and Screensaver lacks the flexibility of Brunn, as the two former are mainly directed towards primary screening where substances are tested in one concentration to find hits or misses according to some threshold value. When doing mechanistic classifications , multiple concentrations need to be assayed and the half maximal inhibitory concentration (IC50) calculated from the accumulated data. In the latter situation, which is the situation Brunn is designed for, the goal is rather to learn more about fewer substances. Moreover, the integration of Brunn with Bioclipse allows the experienced Bioclipse user to seamlessly integrate the Brunn features with a wide plethora of ever growing Bioclipse chemobioinformatics resources and plugins.
Brunn is an open source laboratory information system for microplate based high-throughput screening, capable of handling a multitude of different plate layouts. Brunn can be used to create dose-response curves and extract IC50-values. Brunn uses a point-and-click interface based on the graphical workbench Bioclipse and is easy to use and allows integrating data with the multitude of other resources available in Bioclipse.
Brunn ensures that researchers can retrieve data in a consistent manner regardless of who performed the actual lab work. By constructing platelayouts errors from copying and pasting in spreadsheets can be avoided. The more advanced user can extract Brunn data using the Bioclipse scripting environment.
Although the current implementation is very much tailored for the daily work at the division of Clinical Pharmacology at Uppsala University Hospital, the application contains many more general solutions that allow it to be adopted and reused at other locations. The plugin architecture of Bioclipse provides a great flexibility and enables reuse of components built by other Bioclipse projects. It also makes it easy to build plugins for integrating Brunn with other Bioclipse features.
Basing Brunn on Bioclipse has directly brought in a help system including such as cheat-sheets, and scripting possibilities as well as access to a reporting framework (JasperReports). This means also that it will be straightforward for the experienced user to take advantage of the chemobioinformatics tools in Bioclipse.
Project name: Brunn
Project home page: http://brunn.sourceforge.net
Operating system(s): Platform independent
Programming language: Java
License: Eclipse Public License (EPL)
Restrictions to use by non-academics: none
An Oracle VirtualBox image  based on Ubuntu with complete installation and some test data can be downloaded from http://pele.farmbio.uu.se/trybrunn/BrunnOnUbuntu.zip
The authors would like to thank Jonas Berglund who made the design and a proof-of-concept implementation of the MasterPlate and Plate reports, using the Birt report framework , as well as Carl Mäsak who contributed in the implementation of the parser for the calculation functions. We also want to thank Lena Lenhammar and Kristin Blom for skillfully and patiently testing the software and providing valuable feedback. Financial support to complete this work was obtained from Uppsala University (KoF 07).
- Claus BL, Underwood DJ: Discovery informatics: its evolving role in drug discovery. Drug Discov Today 2002, 7(18):957–966. 10.1016/S1359-6446(02)02433-9View ArticlePubMedGoogle Scholar
- The academic pursuit of screening Nat Chem Biol 2007, 3(8):433–433.Google Scholar
- Brazma A: On the importance of standardisation in life sciences. Bioinformatics 2001, 17(2):113–114. 10.1093/bioinformatics/17.2.113View ArticlePubMedGoogle Scholar
- Schopfer U, Andreae MRM, Hueber M, Saur A, Kummer MO, Girod M, Fox D, Steiner T, Popov M, Smith R: Compound Hub: Efficiency gains through innovative sample management processes. Combinatorial Chemistry and High Throughput Screening 2007, 10(4):283–287. 10.2174/138620707780636655View ArticlePubMedGoogle Scholar
- Quintero C, Kariv I: Design and Implementation of an Automated Compound Management System in Support of Lead Optimization. Journal of Biomolecular Screening 2009, 14(5):499–508. 10.1177/1087057109335326View ArticlePubMedGoogle Scholar
- Spjuth O, Helmus T, Willighagen E, Kuhn S, Eklund M, Wagener J, Murray-Rust P, Steinbeck C, Wikberg J: Bioclipse: an open source workbench for chemo- and bioinformatics. BMC Bioinformatics 2007, 8: 59. 10.1186/1471-2105-8-59PubMed CentralView ArticlePubMedGoogle Scholar
- Astle TW: Recollections of Early Microplate Automation. Journal of the Association for Laboratory Automation 2000, 5(6):30–31. 10.1016/S1535-5535(04)00103-0View ArticleGoogle Scholar
- Spjuth O, Willighagen EL, Guha R, Eklund M, Wikberg JE: Towards interoperable and reproducible QSAR analyses: Exchange of datasets. J Cheminform 2010, 2: 5. 10.1186/1758-2946-2-5PubMed CentralView ArticlePubMedGoogle Scholar
- Carlsson L, Spjuth O, Adams S, Glen RC, Boyer S: Use of historic metabolic biotransformation data as a means of anticipating metabolic sites using MetaPrint2D and Bioclipse. BMC Bioinformatics 2010, 11: 362. 10.1186/1471-2105-11-362PubMed CentralView ArticlePubMedGoogle Scholar
- Willighagen E, Alvarsson J, Andersson A, Eklund M, Lampa S, Lapins M, Spjuth O, Wikberg J: Linking the Resource Description Framework to cheminformatics and proteochemometrics. Journal of Biomedical Semantics 2011, 2(Suppl 1):S6. 10.1186/2041-1480-2-S1-S6PubMed CentralView ArticlePubMedGoogle Scholar
- Wagener J, Spjuth O, Willighagen EL, Wikberg JES: XMPP for cloud computing in bioinformatics supporting discovery and invocation of asynchronous web services. BMC Bioinformatics 2009, 10: 279. 10.1186/1471-2105-10-279PubMed CentralView ArticlePubMedGoogle Scholar
- JasperForge: JasperReports Project Page[http://jasperforge.org/projects/jasperreports]
- Spjuth O, Alvarsson J, Berg A, Eklund M, Kuhn S, Masak C, Torrance G, Wagener J, Willighagen E, Steinbeck C, Wikberg J: Bioclipse 2: A scriptable integration platform for the life sciences. BMC Bioinformatics 2009, 10: 397. 10.1186/1471-2105-10-397PubMed CentralView ArticlePubMedGoogle Scholar
- Hibernate Object Relational Mapping Framework[https://www.hibernate.org/]
- MySQL homepage[http://www.mysql.com/]
- Spring application framework[http://www.springsource.org/]
- Brideau C, Gunter B, Pikounis B, Liaw A: Improved Statistical Methods for Hit Selection in High-Throughput Screening. Journal of Biomolecular Screening 2003, 8(6):634–647. 10.1177/1087057103258285View ArticlePubMedGoogle Scholar
- Commons-Math: The Apache Commons Mathematics Library2011. [http://commons.apache.org/math/]
- Alvarsson J, Andersson C: Performing a double sided T-test on Brunn data.[http://www.myexperiment.org/workflows/1509]
- De Roure D, Goble C, Stevens R: The design and realisation of the my Experiment Virtual Research Environment for social sharing of workflows. Future Generation Computer Systems 2009, 25(5):561–567. 10.1016/j.future.2008.06.010View ArticleGoogle Scholar
- Lindhagen E, Nygren P, Larsson R: The fluorometric microculture cytotoxicity assay. Nat Protocols 2008, 3(8):1364–1369. 10.1038/nprot.2008.114View ArticlePubMedGoogle Scholar
- Kelley BP, Lunn MR, Root DE, Flaherty SP, Martino AM, Stockwell BR: A Flexible Data Analysis Tool for Chemical Genetic Screens. 2004, 11(11):1495–1503.Google Scholar
- Tolopko A, Sullivan J, Erickson S, Wrobel D, Chiang S, Rudnicki K, Rudnicki S, Nale J, Selfors L, Greenhouse D, Muhlich J, Shamu C: Screensaver: an open source lab information management system (LIMS) for high throughput screening facilities. BMC Bioinformatics 2010, 11: 260. 10.1186/1471-2105-11-260PubMed CentralView ArticlePubMedGoogle Scholar
- Dhar S, Nygren P, Csoka K, Botling J, Nilsson K, Larsson R: Anti-cancer drug characterisation using a human cell line panel representing defined types of drug resistance. British journal of cancer 1996, 74(6):888–896. [Cited By (since 1996): 54] [Cited By (since 1996): 54] 10.1038/bjc.1996.453PubMed CentralView ArticlePubMedGoogle Scholar
- Oracle Virtual Box[http://www.virtualbox.org/]
- Eclipse Birt Project[http://eclipse.org/birt/phoenix/]
- Brunn web page[http://brunn.sourceforge.net/]
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.