Clustering Likelihood Curves: Finding Deviations from Single Clusters

Authors:
Claudia Hundertmark;Frank Klawonn
Affiliations:
Department of Cell Biology, Helmholtz Centre for Infection Research, Braunschweig, Germany D-38124;Department of Computer Science, University of Applied Sciences Braunschweig/Wolfenbuettel, Wolfenbuettel, Germany D-38302
Venue:
HAIS '08 Proceedings of the 3rd international workshop on Hybrid Artificial Intelligence Systems
Year:
2008

Citing 2
Cited 1

Fuzzy Models and Algorithms for Pattern Recognition and Image Processing

Fuzzy Models and Algorithms for Pattern Recognition and Image Processing
A Maximum Likelihood Approach to Noise Estimation for Intensity Measurements in Biology

ICDMW '06 Proceedings of the Sixth IEEE International Conference on Data Mining - Workshops

SignalNet: visualization of signal network responses by quantitative proteome data

IDEAL'11 Proceedings of the 12th international conference on Intelligent data engineering and automated learning

Quantified Score

Hi-index	0.00

Visualization

Abstract

For systematic analyses of quantitative mass spectrometry data a method was developed in order to reveal peptides within a protein, that show differences in comparison with the remaining peptides of the protein concerning their regulatory characteristics. Regulatory information is calculated and visualised by a probabilistic approach resulting in likelihood curves. On the other hand the algorithm for the detection of one or more clusters is based on fuzzy clustering, so that our hybrid approach combines probabilistic concepts as well as principles from soft computing. The test is able to decide whether peptides belonging to the same protein, cluster into one or more group. In this way obtained information is very valuable for the detection of single peptides or peptide groups which can be regarded as regulatory outliers.