Cutaneous Leishmaniasis (CL) is a major global health problem requiring appropriate diagnosis methods. Its diagnosis is challenging, particularly in resource-limited settings. The integration of Artificial Intelligence (AI) into medical diagnostics has shown promising results in various fields, including dermatology. In this systematic review, we aim to highlight the value of using AI for CL diagnosis and the AI-based algorithms that are employed in this process, and to identify gaps that need to be addressed. Our work highlights that only a limited number of studies are related to using AI algorithms for CL diagnosis. Among these studies, seven gaps were identified for future research. Addressing these considerations will pave the way for the development of robust AI systems and encourage more research in CL detection by AI. This could contribute to improving CL diagnosis and, ultimately, healthcare outcomes in CL-endemic regions.

In Morocco, cutaneous leishmaniasis (CL) caused by Leishmania (L.) tropica is an important health problem. Despite the high incidence of CL in the country, the genomic heterogeneity of these parasites is still incompletely understood. In this study, we sequenced the genomes of 14 Moroccan isolates of L. tropica collected from confirmed cases of CL to investigate their genomic heterogeneity. Comparative genomics analyses were conducted by applying the recently established Genome Instability Pipeline (GIP), which allowed us to conduct phylogenomic and principal components analyses (PCA), and to assess genomic variations at the levels of the karyotype, gene copy number, single nucleotide polymorphisms (SNPs) and small insertions/deletions (INDELs) variants. Read-depth analyses revealed a mostly disomic karyotype, with the exception of the stable tetrasomy of chromosome 31. In contrast, we identified important gene copy number variations across all isolates, which affect known virulence genes and thus were probably selected in the field. SNP-based cluster analysis of the 14 isolates revealed a core group of 12 strains that formed a tight cluster and shared 45.1 % (87 751) of SNPs, as well as two strains (M3015, Ltr_16) that clustered separately from each other and the core group, suggesting the circulation of genetically highly diverse strains in Morocco. Phylogenetic analysis, which compared our 14 L. tropica isolates against 40 published genomes of L. tropica from a diverse array of locations, confirmed the genetic difference of our Moroccan isolates from all other isolates examined. In conclusion, our results indicate potential regional variations in SNP profiles that may differentiate Moroccan L. tropica from other L. tropica strains circulating in endemic countries in the Middle East. Our report paves the way for future research with a larger number of strains that will allow correlation of diverse phenotypes (resistance to treatments, virulence) and origins (geography, host species, year of isolation) to defined genomic signals such as gene copy number variations or SNP profiles that may represent interesting biomarker candidates

Most graph learning methods remain limited to undirected, pairwise interactions, restricting their ability to capture the multi-entity and directional relationships common in real-world systems. We propose the Directed Higher-Ordered Neural Network (HONN) framework that introduces directionality into hypergraph learning through flexible spectral Laplacian formulations. Unlike fixed-Laplacian methods such as the Generalized Directed Hypergraph Neural Network (GeDi-HNN), a tunable q-parameter in our framework balances local identity preservation with global diffusion, enabling robust and generalizable feature propagation. Experiments on five benchmark datasets show that HONN consistently matches or outperforms state-of-the-art baselines, achieving 84% on NTU-2012, 87.4% on WebKB Texas, and 86.2% on Cornell, while maintaining computational efficiency. Ablation studies confirm the crucial role of Laplacian selection, activation functions, and q-tuning in shaping model performance. By unifying directionality and higher-order reasoning, HONN provides a scalable foundation for predictive modeling in domains such as knowledge graphs, spatio-temporal networks, and recommendation systems.