Page 84 - MI-2-2

Microbes & Immunity

ORIGINAL RESEARCH ARTICLE
Big data and systems biology for biomarker

discovery in chronic hypersensitivity
pneumonitis and drug repurposing using deep

learning-based DTI models

Sung-Yu Lin, Chung-Kuan Liang, and Bor-Sen Chen*
Department of Electrical Engineering, College of Electrical Engineering and Computer Science,
National Tsing Hua University, Hsinchu, Taiwan

Abstract

Chronic hypersensitivity pneumonitis (CHP) is commonly associated with repeated
inhalation of organic dust-containing microbial agents. The identification and
mitigation of exposure to these microbes are key steps in managing and preventing CHP.
Therefore, to elucidate the pathogenetic mechanisms underlying fibrosis and identify
biomarkers for targeted therapeutics, a systems biology approach integrating large-
scale microarray data mining and whole-genome RNA-sequencing data is crucial. In
this study, systems biology methodologies were employed to construct genome-wide
*Corresponding author:
Bor-Sen Chen genetic and epigenetic networks for fibrotic lung cells from CHP patients and healthy
(bschen@ee.nthu.edu.tw) controls. Protein-protein interaction networks and gene/microRNA/long non-coding
Citation: Lin S, Liang C , Chen B. RNA regulatory networks were established using RNA sequencing data, followed by
Big data and systems biology for refinement through a system identification framework. By applying principal network
biomarker discovery in chronic projection and the Kyoto Encyclopedia of Genes and Genomes pathway annotation,
hypersensitivity pneumonitis and
drug repurposing using deep core signaling pathways and their downstream cellular dysfunctions were identified,
learning-based DTI models. revealing key signaling pathways, such as the mitogen-activated protein kinase,
Microbes & Immunity. 2025; phosphoinositide 3-kinase, chemokine (C-X-C motif) ligand (CXCL)/CXC chemokine
2(2):76-105.
doi: 10.36922/mi.4620 receptors family, and chemokine (C-C motif) ligand (CCL) family signaling pathways.
In addition, significant biomarkers of pathogenetic mechanisms of CHP, such as AKT1,
Received: August 21, 2024
CCL20, cell division cycle 23, CXCL1, nuclear factor-kappa B, and tumor necrosis factor,
1st revised: December 11, 2024 were identified as potential drug targets of CHP. Significant biomarkers associated
2nd revised: January 6, 2025 with cell inhibitors, cellular dysfunction, and cellular disorders were identified as
drug targets in the core disease-causing signaling pathways. Finally, using a deep
Accepted: January 20, 2025
neural network-based drug-target interaction model and drug design specification,
Published online: March 28, 2025 three molecular drugs, azathioprine, masitinib, and primaquine, were screened and
Copyright: © 2025 Author(s). proposed as candidates for a multiple-molecular drug for the treatment of CHP.
This is an Open-Access article
distributed under the terms of the
Creative Commons Attribution Keywords: Chronic hypersensitivity pneumonitis; Systems biology; Genome-wide genetic
License, permitting distribution, and epigenetic networks; Deep neural network-based drug-target interaction model
and reproduction in any medium,
provided the original work is
properly cited.
Publisher’s Note: AccScience
Publishing remains neutral with 1. Introduction
regard to jurisdictional claims in
published maps and institutional Chronic hypersensitivity pneumonitis (CHP), also known as chronic bird fancier’s lung
affiliations. or chronic pigeon breeder’s lung, is a complex interstitial lung disease characterized by an

Volume 2 Issue 2 (2025) 76 doi: 10.36922/mi.4620