iOmicsPASS: network-based integration of multiomics data for predictive subnetwork discovery

Hiromi W.L. Koh, Damian Fermin, Christine Vogel, Kwok Pui Choi, Rob M. Ewing, Hyungwon Choi

Research output: Contribution to journalArticle

Abstract

Computational tools for multiomics data integration have usually been designed for unsupervised detection of multiomics features explaining large phenotypic variations. To achieve this, some approaches extract latent signals in heterogeneous data sets from a joint statistical error model, while others use biological networks to propagate differential expression signals and find consensus signatures. However, few approaches directly consider molecular interaction as a data feature, the essential linker between different omics data sets. The increasing availability of genome-scale interactome data connecting different molecular levels motivates a new class of methods to extract interactive signals from multiomics data. Here we developed iOmicsPASS, a tool to search for predictive subnetworks consisting of molecular interactions within and between related omics data types in a supervised analysis setting. Based on user-provided network data and relevant omics data sets, iOmicsPASS computes a score for each molecular interaction, and applies a modified nearest shrunken centroid algorithm to the scores to select densely connected subnetworks that can accurately predict each phenotypic group. iOmicsPASS detects a sparse set of predictive molecular interactions without loss of prediction accuracy compared to alternative methods, and the selected network signature immediately provides mechanistic interpretation of the multiomics profile representing each sample group. Extensive simulation studies demonstrate clear benefit of interaction-level modeling. iOmicsPASS analysis of TCGA/CPTAC breast cancer data also highlights new transcriptional regulatory network underlying the basal-like subtype as positive protein markers, a result not seen through analysis of individual omics data.

Original languageEnglish (US)
Article number22
Journalnpj Systems Biology and Applications
Volume5
Issue number1
DOIs
StatePublished - Dec 1 2019

Fingerprint

Molecular interactions
Gene Regulatory Networks
Statistical Models
Data integration
Joints
Interaction
Genome
Breast Neoplasms
Genes
Availability
Proteins
Signature
Datasets
Error Model
Differential Expression
Regulatory Networks
Biological Networks
Data Integration
Centroid
Breast Cancer

ASJC Scopus subject areas

  • Modeling and Simulation
  • Biochemistry, Genetics and Molecular Biology(all)
  • Drug Discovery
  • Computer Science Applications
  • Applied Mathematics

Cite this

iOmicsPASS : network-based integration of multiomics data for predictive subnetwork discovery. / Koh, Hiromi W.L.; Fermin, Damian; Vogel, Christine; Choi, Kwok Pui; Ewing, Rob M.; Choi, Hyungwon.

In: npj Systems Biology and Applications, Vol. 5, No. 1, 22, 01.12.2019.

Research output: Contribution to journalArticle

Koh, Hiromi W.L. ; Fermin, Damian ; Vogel, Christine ; Choi, Kwok Pui ; Ewing, Rob M. ; Choi, Hyungwon. / iOmicsPASS : network-based integration of multiomics data for predictive subnetwork discovery. In: npj Systems Biology and Applications. 2019 ; Vol. 5, No. 1.
@article{7fb685a346ba4ec6a0cbab04a3914fe0,
title = "iOmicsPASS: network-based integration of multiomics data for predictive subnetwork discovery",
abstract = "Computational tools for multiomics data integration have usually been designed for unsupervised detection of multiomics features explaining large phenotypic variations. To achieve this, some approaches extract latent signals in heterogeneous data sets from a joint statistical error model, while others use biological networks to propagate differential expression signals and find consensus signatures. However, few approaches directly consider molecular interaction as a data feature, the essential linker between different omics data sets. The increasing availability of genome-scale interactome data connecting different molecular levels motivates a new class of methods to extract interactive signals from multiomics data. Here we developed iOmicsPASS, a tool to search for predictive subnetworks consisting of molecular interactions within and between related omics data types in a supervised analysis setting. Based on user-provided network data and relevant omics data sets, iOmicsPASS computes a score for each molecular interaction, and applies a modified nearest shrunken centroid algorithm to the scores to select densely connected subnetworks that can accurately predict each phenotypic group. iOmicsPASS detects a sparse set of predictive molecular interactions without loss of prediction accuracy compared to alternative methods, and the selected network signature immediately provides mechanistic interpretation of the multiomics profile representing each sample group. Extensive simulation studies demonstrate clear benefit of interaction-level modeling. iOmicsPASS analysis of TCGA/CPTAC breast cancer data also highlights new transcriptional regulatory network underlying the basal-like subtype as positive protein markers, a result not seen through analysis of individual omics data.",
author = "Koh, {Hiromi W.L.} and Damian Fermin and Christine Vogel and Choi, {Kwok Pui} and Ewing, {Rob M.} and Hyungwon Choi",
year = "2019",
month = "12",
day = "1",
doi = "10.1038/s41540-019-0099-y",
language = "English (US)",
volume = "5",
journal = "npj Systems Biology and Applications",
issn = "2056-7189",
publisher = "Nature Publishing Group",
number = "1",

}

TY - JOUR

T1 - iOmicsPASS

T2 - network-based integration of multiomics data for predictive subnetwork discovery

AU - Koh, Hiromi W.L.

AU - Fermin, Damian

AU - Vogel, Christine

AU - Choi, Kwok Pui

AU - Ewing, Rob M.

AU - Choi, Hyungwon

PY - 2019/12/1

Y1 - 2019/12/1

N2 - Computational tools for multiomics data integration have usually been designed for unsupervised detection of multiomics features explaining large phenotypic variations. To achieve this, some approaches extract latent signals in heterogeneous data sets from a joint statistical error model, while others use biological networks to propagate differential expression signals and find consensus signatures. However, few approaches directly consider molecular interaction as a data feature, the essential linker between different omics data sets. The increasing availability of genome-scale interactome data connecting different molecular levels motivates a new class of methods to extract interactive signals from multiomics data. Here we developed iOmicsPASS, a tool to search for predictive subnetworks consisting of molecular interactions within and between related omics data types in a supervised analysis setting. Based on user-provided network data and relevant omics data sets, iOmicsPASS computes a score for each molecular interaction, and applies a modified nearest shrunken centroid algorithm to the scores to select densely connected subnetworks that can accurately predict each phenotypic group. iOmicsPASS detects a sparse set of predictive molecular interactions without loss of prediction accuracy compared to alternative methods, and the selected network signature immediately provides mechanistic interpretation of the multiomics profile representing each sample group. Extensive simulation studies demonstrate clear benefit of interaction-level modeling. iOmicsPASS analysis of TCGA/CPTAC breast cancer data also highlights new transcriptional regulatory network underlying the basal-like subtype as positive protein markers, a result not seen through analysis of individual omics data.

AB - Computational tools for multiomics data integration have usually been designed for unsupervised detection of multiomics features explaining large phenotypic variations. To achieve this, some approaches extract latent signals in heterogeneous data sets from a joint statistical error model, while others use biological networks to propagate differential expression signals and find consensus signatures. However, few approaches directly consider molecular interaction as a data feature, the essential linker between different omics data sets. The increasing availability of genome-scale interactome data connecting different molecular levels motivates a new class of methods to extract interactive signals from multiomics data. Here we developed iOmicsPASS, a tool to search for predictive subnetworks consisting of molecular interactions within and between related omics data types in a supervised analysis setting. Based on user-provided network data and relevant omics data sets, iOmicsPASS computes a score for each molecular interaction, and applies a modified nearest shrunken centroid algorithm to the scores to select densely connected subnetworks that can accurately predict each phenotypic group. iOmicsPASS detects a sparse set of predictive molecular interactions without loss of prediction accuracy compared to alternative methods, and the selected network signature immediately provides mechanistic interpretation of the multiomics profile representing each sample group. Extensive simulation studies demonstrate clear benefit of interaction-level modeling. iOmicsPASS analysis of TCGA/CPTAC breast cancer data also highlights new transcriptional regulatory network underlying the basal-like subtype as positive protein markers, a result not seen through analysis of individual omics data.

UR - http://www.scopus.com/inward/record.url?scp=85068765745&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068765745&partnerID=8YFLogxK

U2 - 10.1038/s41540-019-0099-y

DO - 10.1038/s41540-019-0099-y

M3 - Article

C2 - 31312515

AN - SCOPUS:85068765745

VL - 5

JO - npj Systems Biology and Applications

JF - npj Systems Biology and Applications

SN - 2056-7189

IS - 1

M1 - 22

ER -