نوع مقاله : مقاله پژوهشی
عنوان مقاله English
نویسندگان English
Introduction: Lumpy skin disease remains a major economic burden for the cattle industry. By expanding protein functional diversity, alternative splicing can contribute to fine-tuned regulation of host immune and antiviral responses. Here we examined the temporal pattern of alternative splicing in LSDV susceptible and resilient cattle, identified key candidate genes, and asked whether pre-infection molecular differences relate to the distinct clinical outcomes observed in the two groups.
Materials and Methods: Raw single-end RNA-sequencing data were retrieved from the public ArrayExpress repository (accession number E-MTAB-12547), comprising whole-blood samples from five Holstein bulls (three susceptible and two resilient) at four time points: day −5 (pre-infection) and days 3, 7, and 15 post-infections. After quality filtering with Trimmomatic and Cutadapt, and alignment to the Ensembl ARS-UCD2.0 reference genome (release 116) using HISAT2, alternative splicing events were identified with rMATS-turbo. Analyses were performed at two levels: (1) a longitudinal paired design, in which each clinical group was independently compared with its own baseline (six paired tests), consistent with the PAIRADISE statistical framework designed for paired RNA-seq data structures; and (2) a complementary cross-sectional unpaired comparison between the two clinical groups at day −5, to identify background differences independent of infection. Statistical significance criteria were set at FDR < 0.05 and |ΔPSI| ≥ 0.10 (based on JC counts), and technical pseudo-replicates arising from the detection of novel splice sites were removed. Functional enrichment analysis (KEGG and GO) was performed using the DAVID database, protein–protein interaction networks were constructed in STRING, and transcription factors (TFs) were identified using AnimalTFDB (version 4.0). Final candidate genes were visually confirmed with Sashimi plots.
Results and Discussion: The longitudinal analysis identified 57 significant splicing events (23 in the susceptible group, 34 in the resilient group) with no genes shared between groups; most belonged to the RNA-splicing autoregulatory network. The resilient group followed a wave-like pattern splicing response (peaking on day 3, quiescent on day 7, and returning with a focus on DNM2 by day 15) whereas the susceptible group showed a more continuous, broader splicing response that ultimately centered on RAB5A at day 15. Seven final candidate genes were prioritized using a combination of four criteria: effect size (ΔPSI), statistical significance (FDR), persistence across consecutive time points, and pathway specificity; where STRING data were available, network connectivity degree served as an additional interpretive cue. In the resilient group, DNM2 was selected for its specific link to pathogen entry/clearance pathways, and SRSF7 for recurring independently at two time points (days 3 and 15). In the susceptible group, PAPOLA showed the largest effect size in the longitudinal analysis (ΔPSI = 0.822) and is implicated in mRNA surveillance; RPS24 was the only candidate directly linked to a known viral disease (COVID-19); XRN1 was both enriched in the RNA-degradation pathway and emerged as the hub node with the highest connectivity in the day-3 susceptible-group network; HNRNPK, despite a smaller effect size, was retained for its role in the spliceosome pathway and network-level support; and RAB5A was included for its involvement in several pathogen entry/clearance pathways together with a significant day-15 event. Notably, XRN1, HNRNPK, and RAB5A have each been reported elsewhere as host targets exploited by other viruses (including influenza A and foot-and-mouth disease virus) for replication, raising the possibility of an analogous role in LSDV, though this remains speculative. Two genes, APC and SOS2, were statistically very significant but were excluded from the candidate list given their modest effect sizes and annotation across dozens of broad, non-specific pathways, a pattern typical of network hub genes. The PPI analysis lent partial support to this selection. In the susceptible group, significant networks were found on day 3 (16 nodes, 7 edges; P = 0.001) and day 7 (3 nodes, 1 edge; P = 0.03), while the resilient group's day-3 network (21 nodes, 4 edges) did not reach significance (P = 0.07); no edges were detected on day 15 in either group. Within the significant day-3 susceptible network, XRN1, RPS24, and PAPOLA formed a connected subnetwork (independent evidence for their functional relationship beyond enrichment analysis) whereas HNRNPK interacted only with SRSF10. Of the 57 identified genes, seven (ZNF276, ZNF131, TFDP2, MLX, ZNF410, BCLAF1, and NR2C2) were annotated as transcription factors via AnimalTFDB, none overlapping with the seven main candidates. These TF genes were enriched in two GO terms related to RNA polymerase II–mediated transcriptional regulation (5 and 4 genes respectively; Benjamini = 0.0115 and 0.0198). Given the study's limited sample size, however, this should be treated as a preliminary, exploratory pattern rather than firm evidence of a regulatory role.
The complementary day −5 comparison revealed 353 significant events between the two groups, reflecting background molecular differences that predate infection; four of these genes (SRSF7, HNRNPK, XRN1, RAB5A) overlapped with the longitudinal candidates. KEGG enrichment of this gene set highlighted ribosome- and COVID-19–related pathways in the resilient group, and platelet activation, tuberculosis, and nucleocytoplasmic transport pathways in the susceptible group. At the individual-gene level, several innate immunity genes also showed differential pre-infection retention: CLEC4D, IFI47, NMI, and CASP1 in the resilient group, and STAT1, IFI27L2, and CD163 in the susceptible group—a pattern suggesting that baseline differences are concentrated around pathogen-sensing and interferon-signaling components. PPI network significance does not establish causality for any splicing event, relying only on known full-length protein interactions.
Conclusion: This study suggests that alternative splicing may act as a regulatory layer underlying the differential splicing response of cattle to LSDV, with RNA-processing machinery, pathogen entry pathways, and immune signaling all implicated (albeit through different gene sets) in both clinical groups. Pre-infection molecular differences further point to a potential biomarker for predicting susceptibility or resistance. Given the limited sample size (three susceptible versus two resilient animals) and the exploratory nature of this analysis, all findings (candidate genes and enriched pathways alike) should be regarded as preliminary molecular candidates requiring independent validation in larger populations.
کلیدواژهها English
