# Identification of Conserved Gene Expression Signature and Potential Therapeutic Target in Severe Malaria Through Differentially Expressed Genes (DEGs) and Machine Learning Prediction
> Suryandari D.A.

URL kanonis: https://discover.unhas.ac.id/publications/pub_scopus_105043165332
Jurnal / Konferensi: Applied Biosciences
Tahun terbit: 2026
DOI: https://doi.org/10.3390/applbiosci5020049
ISSN: 28130464
Kuartil SJR: Q1
Citations: 0

## Authors
- Suryandari D.A.

## Abstract
Background: Severe malaria remains a major cause of morbidity and mortality, yet the conserved molecular signatures underlying complicated infections across Plasmodium vivax (P. vivax) and Plasmodium falciparum (P. falciparum) are not well characterized. Identifying shared transcriptional biomarkers and host–parasite interaction networks is crucial for improving diagnosis and discovering new therapeutic targets. Methods: Public transcriptomic datasets (GSE55644, GSE59844, GSE34404) were analyzed using GEO2R to identify differentially expressed genes (DEGs). Volcano plots, Venn diagrams, and KEGG mapping were used to identify conserved DEGs. Principal Component Analysis (PCA) and Support Vector Machine (SVM) models were used to assess predictive performance. Host–parasite cross-species correlation analysis integrated parasite DEGs with host hub-genes. Functional enrichment and network module analysis were performed using Cytoscape v3.10.2 and GO/KEGG annotation tools. Results: A total of 3363 DEGs were identified in P. vivax (GSE55644) and only one DEG in P. falciparum (GSE59844) using adjusted p-values, though 772 DEGs emerged with unadjusted p-values. Cross-dataset comparison revealed 18 common DEGs, with eight upregulated genes—TIM9, NUF2, SRP68, HDAC1, GRP94, DHHC8, PPM9, and RPL27—showing robust predictive performance (AUC = 1.000; CA = 1.000) for distinguishing complicated from uncomplicated malaria in both species. Host analysis identified 1719 DEGs and six hub-genes (TNF, IL6, TLR4, CR1, CD40LG, ICAM1) linked to apoptosis, Toll-like receptor signaling, complement cascades, and cell adhesion. SVM validation predicted parasitemia levels with 75.5–84.0% accuracy. Cross-species correlation revealed strong positive interactions between parasite HDAC1/GRP94 and host IL6/TNF and negative correlations involving NUF2, TIM9, ICAM1, and CR1. Functional enrichment analysis highlighted ER stress, immune activation, and erythrocyte adhesion pathways, which together form three major host–parasite modules. Conclusion: These findings highlight conserved biomarkers and potential therapeutic candidates for future validation, demonstrating that combined DEG profiling and machine-learning approaches can provide a powerful framework for improving diagnostics and intervention strategies for severe malaria.

## Keywords
- Biology
- Computational biology
- Transcriptome
- KEGG
- Parasitemia
- Malaria
- Support vector machine
- Gene
- Plasmodium vivax
- Gene expression profiling
- Plasmodium falciparum
- Identification (biology)
- Genetics
- Artificial intelligence
- Machine learning
- Conserved sequence
- Complement (music)
- Principal component analysis
- Proteome
- Gene expression
- Immune system
- Correlation
- Regulation of gene expression
- Bioinformatics

---
Sumber: Discover Unhas — RIMS Universitas Hasanuddin.
Saat mengutip, gunakan DOI bila tersedia atau URL kanonis di atas.