The Comprehensive Systems Biology Project (CSB)
- A Comprehensive Systems-Biology Database -
Hosted at Max Planck Institute of Molecular Plant Physiology
Databases: Associated DB | Transcriptome DB | Metabolome DB | Co-Response DB | BestFit
Home | What's new | Search | About CSB.DB | Site Map | Mail2Us | FAQ | Help
NOTIFICATION: Permanent migration of all CSB.DB services and functionalities on 1.1.2016.
Herewith we inform you that all CSB.DB databases have been migrated at the beginning of 2016. This includes all gene correlation and expression databases, the GMD@CSB.DB module and all associated databases
The BestFit software, a tool for non-aqueous fractionation data analysis, will also be available by the Experimental Systems Biology Research Group headed by Dr. Patrick Giavalisco.
We thank all users, contributors, and collaborators of CSB.DB at the Max Planck Institute for their long-standing support.
Yours sincerely, the CSB.DB Curator and the CSB.DB Developmental Core Team
Most of the information provided by this server are within the public domain and can be accessed via Web interface. The public domain information on the CSB.DB Web pages can be freely downloaded here and reproduced. Please contact the CSB.DB-Curator if you miss necessary information therein.

Below are the list a frequently asked questions.
Frequently Asked Questions (FAQs)
Currently Help information can be accessed by pressing on the information button (i) or by pressing on the detailed information button (iD). If you think you found an error or a missing feature contact us...more help
CSB.DB is meant to help finding transcripts/metabolites/proteins etc. which are co-regulated with other transcripts/metabolites/proteins. Just assume you identified two downstream target genes of your favourite transcription factor. Then if you are lucky they would be co-regulated and now you could ask the question, which other genes are also co-regulated and might thus also be target genes. That is where CSB.DB comes in. In this case an intersection gene query would be what you wanted.
Your screen should have a resolution of (minimal 1024 x 768) and disable the fovorites. You need a JavaScript enabled browser and must allow cookies for full functionality and full help. HTMLs and JavaScripts were tested on MS IE 6.0, Netscape Navigator 7.2, Mozilla 1.4 & Opera 7.2. All HTMLs should run on MS IE for Macintosh 5.2 or higher, MS IE 5.0 or higher, Mozilla 1.0 or higher, Netscape Navigator 7.0 or higher, Opera 6.0 or higher and Safari 1.0 or higher.
For various HTML pages we use png instead of gif files. Not all browser (especially older) support the png files. Update your browser to get the newest stable version, e.g. for Microsoft see
Please use our text exporter on the top of the query result pages. The downloaded file is zipped. Unzip with WinZip or other software and you can upload the text file in MS Excel or in a text editor. The results are tab delimited.
Make sure you have java version 1.4 or better installed (java download) and make sure you enabled java support for the internet. You can test it (incl. version) on Sun's Test Java Virtual Machine web page.
If you want to figure out, if some of your genes cluster together or if special genes form 'hubs' in a response network. If the latter is the case they might represent important regulation points.
Due to the generation of the correlation matrix which discards data spots not well measured, it happens that whole transcript spots are not available. Just imagine the following situation: Six microarray experiments, where in the first three you measure YFG1 but for YFG2 you only get absent calls. In the next three you can measure YFG2 but get absent calls for YFG1. Since correlation is based on genes being CO-regulated no correlation between those two could be measured.
These are correlation coefficients which measure how strongly transcripts/metabolites etc. are co-regulated over experiments. They are a measure how much alike the behaviour is over the experiments, e.g. do they go up and down together. In general Pearson is parametric and the other two are non-parametric making them more robust to individual outliers...more
After querying for your gene click the links on the correlation coefficients between your gene and a second gene. A new page will open, where you can amongst others choose scatter plot.
Mark the whole table press Ctrl-C (Apple-C) in your spreadsheet application press Ctrl-V (Apple-V).
They were assembled by Oliver Thimm at the Max Planck Institute of Molecular Plant Physiology (MapMan; Thimm et al., 2004). The annotations are currently revised.
An obvious resource would be the Stanford Microarray Database (SMD). CSB.DB only incorporates correlation between transcripts/metabolites etc but not the raw values.

For suggestions and questions feel free to contact the CSB.DB curator.
Minimal resolution for optimal view: 1024x768 (without favorites)! Web browser of the sixth generation required, e.g. MS IE 6.0. Javascript must be enabled... more