Computational identification of key functional genes associated with fusarium verticillioides pathogenicity

There is a critical need to better understand the pathogenicity of the fungus Fusarium verticillioides (teleomorph Gibberella moniliformis) leading to ear rots and stalk rots of maize worldwide. The underlying molecular and cellular mechanisms associated with the fungal pathogenicity are complex, and thus a better systematic analysis is needed. In this work, we performed multiple analytical methods to predict key F. verticillioides pathogenicity genes in the subnetwork modules predicted in our recent work. In order to identify potential pathogenicity genes, we first prepared the RNA-seq infection transcriptome and constructed their co-expression networks, and predicted potential subnetwork modules as described in our previous work. Subsequently, we analytically investigated whether each gene in its predicted subnetwork module satisfied several conditions; i) highly impactful in a probabilistic manner, ii) relatively differentially correlated between two strains (wild type vs mutant), iii) relatively more connected in the given module, iv) relatively highly expressed in wild type, v) orthologous to known pathogenic genes in other fungi, and vi) annotated to significant GO terms with other member genes. Through our systematic investigation of the RNA-seq data, we have identified potential F. verticillioides functional genes that demonstrated having not only harmonious coordination with the other genes in a module, but also significant influence on other member genes. Furthermore, the potential genes showed strong differentiation between the two phenotypes and also association with the pathogenicity mechanism.