Novel visualization methods for protein data

S. Mumtaz; I.T. Nabney; D. Flower

doi:10.1109/CIBCB.2012.6217231

Novel visualization methods for protein data

S. Mumtaz, I.T. Nabney, D. Flower

Computing Science

Research output: Chapter in Book/Report/Conference proceeding › Published conference contribution

2 Citations (Scopus)

Abstract

Visualization of high-dimensional data has always been a challenging task. Here we discuss and propose variants of non-linear data projection methods (Generative Topographic Mapping (GTM) and GTM with simultaneous feature saliency (GTM-FS)) that are adapted to be effective on very high-dimensional data. The adaptations use log space values at certain steps of the Expectation Maximization (EM) algorithm and during the visualization process. We have tested the proposed algorithms by visualizing electrostatic potential data for Major Histocompatibility Complex (MHC) class-I proteins. The experiments show that the variation in the original version of GTM and GTM-FS worked successfully with data of more than 2000 dimensions and we compare the results with other linear/nonlinear projection methods: Principal Component Analysis (PCA), Neuroscale (NSC) and Gaussian Process Latent Variable Model (GPLVM).

Original language	English
Title of host publication	2012 IEEE Symposium on Computational Intelligence and Computational Biology, CIBCB 2012
Number of pages	8
ISBN (Electronic)	978-1-4673-1191-5
DOIs	https://doi.org/10.1109/CIBCB.2012.6217231
Publication status	Published - 14 Jun 2012

Access to Document

10.1109/CIBCB.2012.6217231Licence: Unspecified

Cite this

@inproceedings{a95df6160987407da76f814d4c3f8b64,

title = "Novel visualization methods for protein data",

abstract = "Visualization of high-dimensional data has always been a challenging task. Here we discuss and propose variants of non-linear data projection methods (Generative Topographic Mapping (GTM) and GTM with simultaneous feature saliency (GTM-FS)) that are adapted to be effective on very high-dimensional data. The adaptations use log space values at certain steps of the Expectation Maximization (EM) algorithm and during the visualization process. We have tested the proposed algorithms by visualizing electrostatic potential data for Major Histocompatibility Complex (MHC) class-I proteins. The experiments show that the variation in the original version of GTM and GTM-FS worked successfully with data of more than 2000 dimensions and we compare the results with other linear/nonlinear projection methods: Principal Component Analysis (PCA), Neuroscale (NSC) and Gaussian Process Latent Variable Model (GPLVM).",

author = "S. Mumtaz and I.T. Nabney and D. Flower",

year = "2012",

month = jun,

day = "14",

doi = "10.1109/CIBCB.2012.6217231",

language = "English",

isbn = "978-1-4673-1190-8",

booktitle = "2012 IEEE Symposium on Computational Intelligence and Computational Biology, CIBCB 2012",

}

TY - GEN

T1 - Novel visualization methods for protein data

AU - Mumtaz, S.

AU - Nabney, I.T.

AU - Flower, D.

PY - 2012/6/14

Y1 - 2012/6/14

N2 - Visualization of high-dimensional data has always been a challenging task. Here we discuss and propose variants of non-linear data projection methods (Generative Topographic Mapping (GTM) and GTM with simultaneous feature saliency (GTM-FS)) that are adapted to be effective on very high-dimensional data. The adaptations use log space values at certain steps of the Expectation Maximization (EM) algorithm and during the visualization process. We have tested the proposed algorithms by visualizing electrostatic potential data for Major Histocompatibility Complex (MHC) class-I proteins. The experiments show that the variation in the original version of GTM and GTM-FS worked successfully with data of more than 2000 dimensions and we compare the results with other linear/nonlinear projection methods: Principal Component Analysis (PCA), Neuroscale (NSC) and Gaussian Process Latent Variable Model (GPLVM).

AB - Visualization of high-dimensional data has always been a challenging task. Here we discuss and propose variants of non-linear data projection methods (Generative Topographic Mapping (GTM) and GTM with simultaneous feature saliency (GTM-FS)) that are adapted to be effective on very high-dimensional data. The adaptations use log space values at certain steps of the Expectation Maximization (EM) algorithm and during the visualization process. We have tested the proposed algorithms by visualizing electrostatic potential data for Major Histocompatibility Complex (MHC) class-I proteins. The experiments show that the variation in the original version of GTM and GTM-FS worked successfully with data of more than 2000 dimensions and we compare the results with other linear/nonlinear projection methods: Principal Component Analysis (PCA), Neuroscale (NSC) and Gaussian Process Latent Variable Model (GPLVM).

UR - http://www.scopus.com/inward/record.url?eid=2-s2.0-84864071478&partnerID=MN8TOARS

U2 - 10.1109/CIBCB.2012.6217231

DO - 10.1109/CIBCB.2012.6217231

M3 - Published conference contribution

SN - 978-1-4673-1190-8

BT - 2012 IEEE Symposium on Computational Intelligence and Computational Biology, CIBCB 2012

ER -

Novel visualization methods for protein data

Abstract

Access to Document

Other files and links

Fingerprint

Cite this