A network-based approach for identifying cancer causing pathogens

Joseph Hannigan, Suzanne J. Matthews, John K. Wickiser, Paulo Shakarian

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We present a new method to identify malignant cancer-causing pathogens by analyzing their interactions with the host protein interaction network. We introduce two new measurements, core score and moment score that is based on topological characteristics of the network of host proteins that interact with the pathogen. We applied these measurement to a data set consisting of the interactions of 135 pathogens and a human protein-interaction network. We show a strong linear relationship (R2 = 0:90) between the core score and the probability that a pathogen leads to malignant cancer in humans and demonstrate, using a decision tree classifier, that both measurements can be used to correctly identify pathogens that lead to malignant cancer in humans with an accuracy of 97%.

Original languageEnglish (US)
Title of host publicationProceedings of the 2014 ACM Southeast Regional Conference, ACM SE 2014
PublisherAssociation for Computing Machinery, Inc
ISBN (Print)9781450329231
DOIs
StatePublished - Mar 28 2014
Externally publishedYes
Event2014 ACM Southeast Regional Conference, ACM SE 2014 - Kennesaw, United States
Duration: Mar 28 2014Mar 29 2014

Other

Other2014 ACM Southeast Regional Conference, ACM SE 2014
CountryUnited States
CityKennesaw
Period3/28/143/29/14

Fingerprint

Pathogens
Proteins
Decision trees
Classifiers

Keywords

  • Cancer
  • Network topology
  • Pathogens
  • Protein interaction networks

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Computer Science Applications
  • Hardware and Architecture
  • Software

Cite this

Hannigan, J., Matthews, S. J., Wickiser, J. K., & Shakarian, P. (2014). A network-based approach for identifying cancer causing pathogens. In Proceedings of the 2014 ACM Southeast Regional Conference, ACM SE 2014 [2735459] Association for Computing Machinery, Inc. https://doi.org/10.1145/2638404.2735459

A network-based approach for identifying cancer causing pathogens. / Hannigan, Joseph; Matthews, Suzanne J.; Wickiser, John K.; Shakarian, Paulo.

Proceedings of the 2014 ACM Southeast Regional Conference, ACM SE 2014. Association for Computing Machinery, Inc, 2014. 2735459.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Hannigan, J, Matthews, SJ, Wickiser, JK & Shakarian, P 2014, A network-based approach for identifying cancer causing pathogens. in Proceedings of the 2014 ACM Southeast Regional Conference, ACM SE 2014., 2735459, Association for Computing Machinery, Inc, 2014 ACM Southeast Regional Conference, ACM SE 2014, Kennesaw, United States, 3/28/14. https://doi.org/10.1145/2638404.2735459
Hannigan J, Matthews SJ, Wickiser JK, Shakarian P. A network-based approach for identifying cancer causing pathogens. In Proceedings of the 2014 ACM Southeast Regional Conference, ACM SE 2014. Association for Computing Machinery, Inc. 2014. 2735459 https://doi.org/10.1145/2638404.2735459
Hannigan, Joseph ; Matthews, Suzanne J. ; Wickiser, John K. ; Shakarian, Paulo. / A network-based approach for identifying cancer causing pathogens. Proceedings of the 2014 ACM Southeast Regional Conference, ACM SE 2014. Association for Computing Machinery, Inc, 2014.
@inproceedings{e0c81c42b71c478582d3a8fb7b7e61f3,
title = "A network-based approach for identifying cancer causing pathogens",
abstract = "We present a new method to identify malignant cancer-causing pathogens by analyzing their interactions with the host protein interaction network. We introduce two new measurements, core score and moment score that is based on topological characteristics of the network of host proteins that interact with the pathogen. We applied these measurement to a data set consisting of the interactions of 135 pathogens and a human protein-interaction network. We show a strong linear relationship (R2 = 0:90) between the core score and the probability that a pathogen leads to malignant cancer in humans and demonstrate, using a decision tree classifier, that both measurements can be used to correctly identify pathogens that lead to malignant cancer in humans with an accuracy of 97{\%}.",
keywords = "Cancer, Network topology, Pathogens, Protein interaction networks",
author = "Joseph Hannigan and Matthews, {Suzanne J.} and Wickiser, {John K.} and Paulo Shakarian",
year = "2014",
month = "3",
day = "28",
doi = "10.1145/2638404.2735459",
language = "English (US)",
isbn = "9781450329231",
booktitle = "Proceedings of the 2014 ACM Southeast Regional Conference, ACM SE 2014",
publisher = "Association for Computing Machinery, Inc",

}

TY - GEN

T1 - A network-based approach for identifying cancer causing pathogens

AU - Hannigan, Joseph

AU - Matthews, Suzanne J.

AU - Wickiser, John K.

AU - Shakarian, Paulo

PY - 2014/3/28

Y1 - 2014/3/28

N2 - We present a new method to identify malignant cancer-causing pathogens by analyzing their interactions with the host protein interaction network. We introduce two new measurements, core score and moment score that is based on topological characteristics of the network of host proteins that interact with the pathogen. We applied these measurement to a data set consisting of the interactions of 135 pathogens and a human protein-interaction network. We show a strong linear relationship (R2 = 0:90) between the core score and the probability that a pathogen leads to malignant cancer in humans and demonstrate, using a decision tree classifier, that both measurements can be used to correctly identify pathogens that lead to malignant cancer in humans with an accuracy of 97%.

AB - We present a new method to identify malignant cancer-causing pathogens by analyzing their interactions with the host protein interaction network. We introduce two new measurements, core score and moment score that is based on topological characteristics of the network of host proteins that interact with the pathogen. We applied these measurement to a data set consisting of the interactions of 135 pathogens and a human protein-interaction network. We show a strong linear relationship (R2 = 0:90) between the core score and the probability that a pathogen leads to malignant cancer in humans and demonstrate, using a decision tree classifier, that both measurements can be used to correctly identify pathogens that lead to malignant cancer in humans with an accuracy of 97%.

KW - Cancer

KW - Network topology

KW - Pathogens

KW - Protein interaction networks

UR - http://www.scopus.com/inward/record.url?scp=84940688901&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84940688901&partnerID=8YFLogxK

U2 - 10.1145/2638404.2735459

DO - 10.1145/2638404.2735459

M3 - Conference contribution

AN - SCOPUS:84940688901

SN - 9781450329231

BT - Proceedings of the 2014 ACM Southeast Regional Conference, ACM SE 2014

PB - Association for Computing Machinery, Inc

ER -