Insect pathogens are a valuable tool to manage crop pests. They can be used in combination with insecticides to improve their performance and reduce the occurrence of resistance to commercially active ingredients The genus Metarhizium (Ascomycota: Hypocreales) is highly abundant, globally distributed, and best known for infecting and killing many different arthropods. Some Metarhizium species have a broad host range (generalists). In contrast, others show specificity for certain insect families (specialists) and can be used to test hypotheses regarding speciation and host specificity. In this work two Metarhizium species were investigated, M. anisopliae and M. rilyei, and their propagules produced in vitro: blastospores and aerial conidia. Many studies have shown that sequence data can provide crucial information about how these organisms reproduce and persist in different environments. Using Metarhizium spp. as a model, the first chapter briefly present the entomopathogenic fungi infection process and the benefits of using conidia and blastospores for biological pest control in the field. It also shows the main questions and hypotheses that build this work and introduce the advances achieved. The second chapter, a review article, highlights recent advances in genomics and molecular biology of entomopathogenic fungi to further investigate the gene-for-gene relationships in insect-fungus interactions. In the third chapter we investigated the biological and genetic factors that allow blastospores to infect insects and make them potentially effective for biological control in the field. While the generalist M. anisopliae produces under high osmolarity conditions smaller blastospores in more significant number, the Lepidopteran specialist M. rileyi produces fewer propagules with a higher cell volume. We also compared the virulence of blastospores and conidia of these two Metarhizium species towards the economically important caterpillar pest Spodoptera frugiperda. We found that conidia and blastospores from M. anisopliae were both infectious but killed fewer insects compared with M. rileyi conidia and blastospores, where M. rileyi conidia had the highest virulence. Using comparative transcriptomics during propagule penetration on insect cuticles, this third chapter also shows that M. rileyi blastospores expresses more virulence-related genes against S. frugiperda than M. anisopliae. In contrast, conidia of both fungi express more virulence-related oxidative stress factors than blastospores. These results highlight that blastospores use a different virulence mechanism than conidia, which may be explored in new biological control strategies. Finally, chapter four investigated the phenotypical plasticity of M. anisopliae blastospores among different insect species. We determined that M. anisopliae blastospores are highly virulent for Tenebrio molitor, and the percentage of appressorium formation in the membranous wings of this insect was three times higher than in the wings of the other insects. We showed a clear difference in the gene expression pattern in the blastospores of M. anisopliae during the infective process in Gryllus assimillis, Spodoptera frugiperda, Apis mellifera, and Tenebrio molitor. This implies that M. anisopliae transcriptome and virulence change remarkably according to the insect, with the most significant differences for G. assimillis. These differences are associated with the expression of enzymes such as proteases, cutinases, lipases, and peptidases which might be related to the degradation of specific compounds of each insect wing; hydrophobins and destruxins which are associated with virulence and secondary metabolites. The gene expression profiles of fungal propagules during penetration on different insect cuticles were characterized to increase our understanding of the fungal process of insect pathogenicity. The results obtained here can potentially help to find gene candidates whose manipulation can ultimately lead to discovering more effective biological control agents.