Multi-Scale Clustering by Building a Robust and Self Correcting Ultrametric Topology on Data Points

Hsieh Fushing, Hui Wang, Kimberly VanderWaal, Brenda Mccowan, Patrice Koehl

Research output: Contribution to journalArticle

21 Citations (Scopus)

Abstract

The advent of high-throughput technologies and the concurrent advances in information sciences have led to an explosion in size and complexity of the data sets collected in biological sciences. The biggest challenge today is to assimilate this wealth of information into a conceptual framework that will help us decipher biological functions. A large and complex collection of data, usually called a data cloud, naturally embeds multi-scale characteristics and features, generically termed geometry. Understanding this geometry is the foundation for extracting knowledge from data. We have developed a new methodology, called data cloud geometry-tree (DCG-tree), to resolve this challenge. This new procedure has two main features that are keys to its success. Firstly, it derives from the empirical similarity measurements a hierarchy of clustering configurations that captures the geometric structure of the data. This hierarchy is then transformed into an ultrametric space, which is then represented via an ultrametric tree or a Parisi matrix. Secondly, it has a built-in mechanism for self-correcting clustering membership across different tree levels. We have compared the trees generated with this new algorithm to equivalent trees derived with the standard Hierarchical Clustering method on simulated as well as real data clouds from fMRI brain connectivity studies, cancer genomics, giraffe social networks, and Lewis Carroll's Doublets network. In each of these cases, we have shown that the DCG trees are more robust and less sensitive to measurement errors, and that they provide a better quantification of the multi-scale geometric structures of the data. As such, DCG-tree is an effective tool for analyzing complex biological data sets.

Original languageEnglish (US)
Article numbere56259
JournalPLoS One
Volume8
Issue number2
DOIs
StatePublished - Feb 12 2013

Fingerprint

topology
Cluster Analysis
Topology
Geometry
Information science
information science
Information Science
Giraffa camelopardalis
Measurement errors
social networks
Biological Science Disciplines
Explosions
explosions
Brain
Genomics
Social Support
Throughput
Magnetic Resonance Imaging
Technology
brain

ASJC Scopus subject areas

  • Agricultural and Biological Sciences(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Medicine(all)

Cite this

Multi-Scale Clustering by Building a Robust and Self Correcting Ultrametric Topology on Data Points. / Fushing, Hsieh; Wang, Hui; VanderWaal, Kimberly; Mccowan, Brenda; Koehl, Patrice.

In: PLoS One, Vol. 8, No. 2, e56259, 12.02.2013.

Research output: Contribution to journalArticle

Fushing, Hsieh ; Wang, Hui ; VanderWaal, Kimberly ; Mccowan, Brenda ; Koehl, Patrice. / Multi-Scale Clustering by Building a Robust and Self Correcting Ultrametric Topology on Data Points. In: PLoS One. 2013 ; Vol. 8, No. 2.
@article{24e7f68d4d2148f48106e85103c86ca2,
title = "Multi-Scale Clustering by Building a Robust and Self Correcting Ultrametric Topology on Data Points",
abstract = "The advent of high-throughput technologies and the concurrent advances in information sciences have led to an explosion in size and complexity of the data sets collected in biological sciences. The biggest challenge today is to assimilate this wealth of information into a conceptual framework that will help us decipher biological functions. A large and complex collection of data, usually called a data cloud, naturally embeds multi-scale characteristics and features, generically termed geometry. Understanding this geometry is the foundation for extracting knowledge from data. We have developed a new methodology, called data cloud geometry-tree (DCG-tree), to resolve this challenge. This new procedure has two main features that are keys to its success. Firstly, it derives from the empirical similarity measurements a hierarchy of clustering configurations that captures the geometric structure of the data. This hierarchy is then transformed into an ultrametric space, which is then represented via an ultrametric tree or a Parisi matrix. Secondly, it has a built-in mechanism for self-correcting clustering membership across different tree levels. We have compared the trees generated with this new algorithm to equivalent trees derived with the standard Hierarchical Clustering method on simulated as well as real data clouds from fMRI brain connectivity studies, cancer genomics, giraffe social networks, and Lewis Carroll's Doublets network. In each of these cases, we have shown that the DCG trees are more robust and less sensitive to measurement errors, and that they provide a better quantification of the multi-scale geometric structures of the data. As such, DCG-tree is an effective tool for analyzing complex biological data sets.",
author = "Hsieh Fushing and Hui Wang and Kimberly VanderWaal and Brenda Mccowan and Patrice Koehl",
year = "2013",
month = "2",
day = "12",
doi = "10.1371/journal.pone.0056259",
language = "English (US)",
volume = "8",
journal = "PLoS One",
issn = "1932-6203",
publisher = "Public Library of Science",
number = "2",

}

TY - JOUR

T1 - Multi-Scale Clustering by Building a Robust and Self Correcting Ultrametric Topology on Data Points

AU - Fushing, Hsieh

AU - Wang, Hui

AU - VanderWaal, Kimberly

AU - Mccowan, Brenda

AU - Koehl, Patrice

PY - 2013/2/12

Y1 - 2013/2/12

N2 - The advent of high-throughput technologies and the concurrent advances in information sciences have led to an explosion in size and complexity of the data sets collected in biological sciences. The biggest challenge today is to assimilate this wealth of information into a conceptual framework that will help us decipher biological functions. A large and complex collection of data, usually called a data cloud, naturally embeds multi-scale characteristics and features, generically termed geometry. Understanding this geometry is the foundation for extracting knowledge from data. We have developed a new methodology, called data cloud geometry-tree (DCG-tree), to resolve this challenge. This new procedure has two main features that are keys to its success. Firstly, it derives from the empirical similarity measurements a hierarchy of clustering configurations that captures the geometric structure of the data. This hierarchy is then transformed into an ultrametric space, which is then represented via an ultrametric tree or a Parisi matrix. Secondly, it has a built-in mechanism for self-correcting clustering membership across different tree levels. We have compared the trees generated with this new algorithm to equivalent trees derived with the standard Hierarchical Clustering method on simulated as well as real data clouds from fMRI brain connectivity studies, cancer genomics, giraffe social networks, and Lewis Carroll's Doublets network. In each of these cases, we have shown that the DCG trees are more robust and less sensitive to measurement errors, and that they provide a better quantification of the multi-scale geometric structures of the data. As such, DCG-tree is an effective tool for analyzing complex biological data sets.

AB - The advent of high-throughput technologies and the concurrent advances in information sciences have led to an explosion in size and complexity of the data sets collected in biological sciences. The biggest challenge today is to assimilate this wealth of information into a conceptual framework that will help us decipher biological functions. A large and complex collection of data, usually called a data cloud, naturally embeds multi-scale characteristics and features, generically termed geometry. Understanding this geometry is the foundation for extracting knowledge from data. We have developed a new methodology, called data cloud geometry-tree (DCG-tree), to resolve this challenge. This new procedure has two main features that are keys to its success. Firstly, it derives from the empirical similarity measurements a hierarchy of clustering configurations that captures the geometric structure of the data. This hierarchy is then transformed into an ultrametric space, which is then represented via an ultrametric tree or a Parisi matrix. Secondly, it has a built-in mechanism for self-correcting clustering membership across different tree levels. We have compared the trees generated with this new algorithm to equivalent trees derived with the standard Hierarchical Clustering method on simulated as well as real data clouds from fMRI brain connectivity studies, cancer genomics, giraffe social networks, and Lewis Carroll's Doublets network. In each of these cases, we have shown that the DCG trees are more robust and less sensitive to measurement errors, and that they provide a better quantification of the multi-scale geometric structures of the data. As such, DCG-tree is an effective tool for analyzing complex biological data sets.

UR - http://www.scopus.com/inward/record.url?scp=84873721756&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84873721756&partnerID=8YFLogxK

U2 - 10.1371/journal.pone.0056259

DO - 10.1371/journal.pone.0056259

M3 - Article

C2 - 23424653

AN - SCOPUS:84873721756

VL - 8

JO - PLoS One

JF - PLoS One

SN - 1932-6203

IS - 2

M1 - e56259

ER -