[LOGO]

2D Gel Matching


Point-Pattern-Matching for the analysis of 2D gel images




Short Overview
Project Description
Members
Links
CAROL program

This project is funded by the DFG. It is a joint project of the department of computer science of the Free University of Berlin and the German Heart Center Berlin. The project is about 2-dimensional gel images which are produced by high resolution gel elctrophoresis techniques.

The left picture is a 2-dimensional gel image, on which a spot-detection algorithm has been run to detect the coordinaates of the points ("spots"), which have been marked with red crosses in the right picture. Each spot represents a single protein. The analysis of those images helps to detect molecular as well as genetical causes of heart deseases.

Up to now the analysis of gel images is based upon the utilization of commercial gel analysis systems (e.g. PDQUEST, MELANIE etc.), which however need manual, time consuming interaction by experienced personnel.

One goal of the project is to develop and implement new algorithms for an effective gel analysis, particularly for the gel matching. Additionally a 2D-gel protein database is going to be build up and maintained. The gel matching routines will be able to compare gel images which have been produced in one laboratory (intra-laboratory comparison) as well as gel images from different laboratories (inter-laboratory comparison), which are widely available via the World Wide Web. The programs to be developed will be accessible via the internet in order to perform a "matching on the net".

The gel matching is one of the essential and also most time consuming requirements for a quantitative and qualitative data analysis of protein gel images. The corresponding algorithmic problem is a 2-dimensional pattern matching problem which has to take into account geometrical distortions that arise during the production of the gel images. The goal is to develop new algorithms for the gel matching based on point pattern matching techniques which are used in computational geometry. Due to the fundamental task of approximate point pattern matching, every progress which is made concerning gel matching techniques will be important for several other applications.


Zu Deutsch wechseln


Last change: November 15th, 1999
These pages have been created by Carola Wenk, wenk@inf.fu-berlin.de
Institute of Computer Science, Takustr. 9, 14195 Berlin, Free University Berlin