SimCluster: pat SimCluster: pat attern recognition attern - PowerPoint PPT Presentation

Universidade de São Paulo Departamento de Computação o e Matemática LabPIB - Laboratório de Processa cessamento de Informação Biológica Prof. Dr. Vêncio rvencio@usp.br SimCluster: pat SimCluster: pat attern recognition attern recognition for composition nal data

Pattern recognition What do we mean by Pattern Recogn gnition ?

Pattern recognition

Pattern recognition

Pattern recognition

Pattern recognition

Biological motivation Joyce & Palsson, Nat Rev Mol Cell Biol 2006

Biological motivation Joyce & Palsson, Nat Rev Mol Cell Biol 2006

Technological tool to be used: quantit titative sequencing

Technological tool to be used: quantit titative sequencing http://www.genome.gov/sequencingcosts/

Next-gen sequencing Nature Reviews Genetics, 2009

The technical problem we are addres ssing Wang et al. , Nature Reviews Genetics, 2009 9

2003

Model – the urn model

2004

2007

Crucial element in clustering analysis is: distance between objects

Pattern recognition

Model

Crucial element in clustering analysis is: distance between objects

Metric properties

Additional “good” properties for comp positional analysis

Additional “good” properties for comp positional analysis

Additional “good” properties for comp positional analysis

Aitchisonean Distance

Aitchisonean Distance

2007

Model validation: simulated sequencin cing data from Affymetrix data

Model validation: simulated sequencin cing data from Affymetrix data

Model validation: simulated sequencin cing data from Affymetrix data - mouse macrophages RNA - stimulated by different Toll-like rec ceptor agonists: LPS, PIC, CPG, R848 and PAM - time-course: 0, 20, 40, 60, 80 and 1 120 minutes. - Gilchrist M, et al . Systems biology approaches ident ntify ATF3 as a negative regulator of Toll-like recep eptor 4. Nature 2006, 441:173–178. - Innate Immunity Systems Biology http://www.innateimmunity-system msbiology.org - transcript abundances proportiona al to Affy signal - simulated sampling using real-data ta structure

Model validation: simulated sequencin cing data from Affymetrix data

circumstantial” evidence) � Model validation: other examples (“cir

Web site freely available for all

Opportunities: there is more in Pattern rn Recognition beyond clustering

Opportunities: there is more in Pattern rn Recognition beyond clustering

Opportunities: there is more in Pattern rn Recognition beyond clustering

Acknowledgements Prof. Carlinhos Pro of. Edson Amaro Jr. IME-USP e BIOINFO-USP FM- -USP e HIAE ������������������������� ����