Vanja Tatić1*, Éva Schád2, Miloš Avramov1, Teodora Knežić Avramov3, Ágnes Révész4, László Drahos4, Ágnes Tantos5 and Željko D. Popović1
1University of Novi Sad, Faculty of Sciences, Department of Biology and Ecology, Novi Sad, Serbia
2HUN-REN Research Centre for Natural Sciences, Institute of Molecular Life Sciences
3University of Novi Sad, BioSense Institute, Serbia
4Institute of Organic Chemistry, Research Centre for Natural Sciences, 1117 Budapest, Hungary
5HUN-REN Research Centre for Natural Sciences
vanja.tatic [at] dbe.uns.ac.rs
Abstract
Proteomic analyses based on mass spectrometry (MS) are widely used to study biological systems, but most available bioinformatic tools are optimized for well-annotated model organisms such as mice, humans, Drosophila or Arabidopsis. Consequently, research on non-model species is hindered by incomplete protein annotation, limited pathway databases, and reduced accuracy in functional enrichment analyses. These limitations restrict the interpretation of MS data and reduce reproducibility across studies. Therefore, there is a growing need for novel and flexible analytical approaches that can accommodate poorly annotated genomes and species-specific proteins. Developing improved computational pipelines and expanding annotation resources would enhance proteomic analysis in non-model organisms, enabling more accurate biological interpretation and supporting broader applications in ecology, agriculture, and evolutionary biology.
Mass spectrometry data from Ostrinia nubilalis larvae infected with two bacterial strains, P. aeruginosa and S. aureus, was analysed with the goal of obtaining a functional annotation of differentially represented proteins in four tissues: epidermis, intestine, fat body and haemolymph. Proteins that showed two-fold increase compared to the control were considered overrepresented, while those that were reduced to at least half were considered underrepresented in each tissue, under each infection treatment. EggNOG-mapper and its database of ortholog groups was used to obtain predicted protein names, KEGG pathways, modules, as well as orthologs and Gene Ontology (GO) labels. The standard for GO enrichment in the R environment, clusterProfilier, was shown to be inadequate for non-model species, so downstream GO analysis was performed using ontologyIndex together with tidyverse. GO annotations were filtered to remove obsolete and unannotated terms, and mapped to broader GO Slim categories using the current Gene Ontology hierarchy. This allowed proteins to be grouped into three major groups: biological processes, molecular functions, and cellular components, and their relative representation across tissues and treatments was quantified and visualised. Overall, this approach provided a simplified functional overview of protein changes in a non-model insect system, making the biological interpretation of large proteomic datasets more accessible.
Keywords: Proteomics; Mass spectrometry; Non-model organisms
Acknowledgement: Acknowledgement: The authors gratefully acknowledge the financial support of the Ministry of Science, Technological Development and Innovation of the Republic of Serbia (Grants No. 451-03-33/2026-03/ 200125, 451-03-34/2026-03/ 200125 and 451-03-33/2026-03/ 200358), the Provincial Secretariat for Higher Education and Scientific Research (Grants No. 003801787 2025 09418 003 000 000 001/1 and No. 003956066 2025 09418 003 000 000 001 04 003) and the Serbian-Hungarian joint research project ''Fighting against antibiotic resistance – identification of prospective antimicrobial peptides from alternative sources (ALARM)''.