Petromyzon marinus HOX Genes Manual Curation GFF

Analysis NamePetromyzon marinus HOX Genes Manual Curation GFF
MethodApollo (2.0.8)
Date performed2017-12-14

Manual curation of HOX Clusters: Manual curation of gene models was carried out using Apollo implemented in JBrowse. Indels in the assembly were identified and corrected by comparison with RNAseq and genomic DNA re-sequencing data. Gene predictions from Maker were refined based on whole embryo RNA-seq data from multiple developmental stages and homology with gene sequences from other vertebrates. In addition to the 42 clustered Hox genes in the genome assembly, 6 further Hox genes were predicted that did not fall within the 6 HOX clusters. To investigate these genes further, the genomic scaffolds harboring these gene loci were extracted and used as queries for alignment against the assembly by BLAST. Five of these gene loci (homologs of hoxA3, D8, C9, B13 and B13a) were found to align with high sequence similarity (>97% identity) across long stretches of their sequence (>4kb, containing predicted Hox coding sequence and flanking, non-coding sequence) to loci of individual members of the 42 clustered lamprey Hox genes. These loci could represent either recent duplications of Hox loci or could be assembly artifacts arising from the relatively high heterozygosity of the lamprey genome. Based on their exceptionally high levels of coding and non-coding sequence similarity to clustered Hox loci, we infer that these 5 loci are assembly artifacts due to polymorphism and that they do not represent additional singleton Hox genes in the lamprey genome. The 6th predicted singleton Hox gene shows equal levels of homology to ANTP-class homeobox genes of both Hox and non-Hox families, suggesting it is a derived ANTP-class homeobox gene and not necessarily a Hox gene.