Research ExplorerPublicationsPatient-level proteomic network prediction by explainable ar...

Patient-level proteomic network prediction by explainable artificial intelligence

Standard

Patient-level proteomic network prediction by explainable artificial intelligence. / Keyl, Philipp; Michael, Bockmayr; Heim, Daniel; Dernbach, Gabriel; Montavon, Grégoire; Müller, Klaus-Robert; Klauschen, Frederick.

In: NPJ PRECIS ONCOL, Vol. 6, No. 1, 35, 07.06.2022.

Research output: SCORING: Contribution to journal › SCORING: Journal article › Research › peer-review

Harvard

Keyl, P, Michael, B, Heim, D, Dernbach, G, Montavon, G, Müller, K-R & Klauschen, F 2022, 'Patient-level proteomic network prediction by explainable artificial intelligence', NPJ PRECIS ONCOL, vol. 6, no. 1, 35. https://doi.org/10.1038/s41698-022-00278-4

APA

Keyl, P., Michael, B., Heim, D., Dernbach, G., Montavon, G., Müller, K-R., & Klauschen, F. (2022). Patient-level proteomic network prediction by explainable artificial intelligence. NPJ PRECIS ONCOL, 6(1), [35]. https://doi.org/10.1038/s41698-022-00278-4

Vancouver

Keyl P, Michael B, Heim D, Dernbach G, Montavon G, Müller K-R et al. Patient-level proteomic network prediction by explainable artificial intelligence. NPJ PRECIS ONCOL. 2022 Jun 7;6(1). 35. https://doi.org/10.1038/s41698-022-00278-4

Bibtex

@article{9b140855b9304354862e49160f7b0b1a,
title = "Patient-level proteomic network prediction by explainable artificial intelligence",
abstract = "Understanding the pathological properties of dysregulated protein networks in individual patients' tumors is the basis for precision therapy. Functional experiments are commonly used, but cover only parts of the oncogenic signaling networks, whereas methods that reconstruct networks from omics data usually only predict average network features across tumors. Here, we show that the explainable AI method layer-wise relevance propagation (LRP) can infer protein interaction networks for individual patients from proteomic profiling data. LRP reconstructs average and individual interaction networks with an AUC of 0.99 and 0.93, respectively, and outperforms state-of-the-art network prediction methods for individual tumors. Using data from The Cancer Proteome Atlas, we identify known and potentially novel oncogenic network features, among which some are cancer-type specific and show only minor variation among patients, while others are present across certain tumor types but differ among individual patients. Our approach may therefore support predictive diagnostics in precision oncology by inferring {"}patient-level{"} oncogenic mechanisms.",
author = "Philipp Keyl and Bockmayr Michael and Daniel Heim and Gabriel Dernbach and Gr{\'e}goire Montavon and Klaus-Robert M{\"u}ller and Frederick Klauschen",
note = "{\textcopyright} 2022. The Author(s).",
year = "2022",
month = jun,
day = "7",
doi = "10.1038/s41698-022-00278-4",
language = "English",
volume = "6",
journal = "NPJ PRECIS ONCOL",
issn = "2397-768X",
publisher = "Springer Nature",
number = "1",

}

RIS

TY - JOUR

T1 - Patient-level proteomic network prediction by explainable artificial intelligence

AU - Keyl, Philipp

AU - Michael, Bockmayr

AU - Heim, Daniel

AU - Dernbach, Gabriel

AU - Montavon, Grégoire

AU - Müller, Klaus-Robert

AU - Klauschen, Frederick

N1 - © 2022. The Author(s).

PY - 2022/6/7

Y1 - 2022/6/7

N2 - Understanding the pathological properties of dysregulated protein networks in individual patients' tumors is the basis for precision therapy. Functional experiments are commonly used, but cover only parts of the oncogenic signaling networks, whereas methods that reconstruct networks from omics data usually only predict average network features across tumors. Here, we show that the explainable AI method layer-wise relevance propagation (LRP) can infer protein interaction networks for individual patients from proteomic profiling data. LRP reconstructs average and individual interaction networks with an AUC of 0.99 and 0.93, respectively, and outperforms state-of-the-art network prediction methods for individual tumors. Using data from The Cancer Proteome Atlas, we identify known and potentially novel oncogenic network features, among which some are cancer-type specific and show only minor variation among patients, while others are present across certain tumor types but differ among individual patients. Our approach may therefore support predictive diagnostics in precision oncology by inferring "patient-level" oncogenic mechanisms.

AB - Understanding the pathological properties of dysregulated protein networks in individual patients' tumors is the basis for precision therapy. Functional experiments are commonly used, but cover only parts of the oncogenic signaling networks, whereas methods that reconstruct networks from omics data usually only predict average network features across tumors. Here, we show that the explainable AI method layer-wise relevance propagation (LRP) can infer protein interaction networks for individual patients from proteomic profiling data. LRP reconstructs average and individual interaction networks with an AUC of 0.99 and 0.93, respectively, and outperforms state-of-the-art network prediction methods for individual tumors. Using data from The Cancer Proteome Atlas, we identify known and potentially novel oncogenic network features, among which some are cancer-type specific and show only minor variation among patients, while others are present across certain tumor types but differ among individual patients. Our approach may therefore support predictive diagnostics in precision oncology by inferring "patient-level" oncogenic mechanisms.

U2 - 10.1038/s41698-022-00278-4

DO - 10.1038/s41698-022-00278-4

M3 - SCORING: Journal article

C2 - 35672443

VL - 6

JO - NPJ PRECIS ONCOL

JF - NPJ PRECIS ONCOL

SN - 2397-768X

IS - 1

M1 - 35

ER -