Neospora caninum is an apicomplexan parasite which can cause abortion in cattle instigating major economic burden. Vaccination has been proposed as the most cost-effective control measure to alleviate this burden. Consequently the overriding aspiration for N. caninum research is the identification and subsequent evaluation of vaccine candidates in animal models. To save on time, cost and effort it is now feasible to use an in silico approach for vaccine candidate prediction. Precise protein sequences, derived from the correct open reading frame, are paramount and arguably the most important factor determining success or failure of the approach. The challenge is that publicly available N. caninum sequences are mostly derived from gene predictions. Annotated inaccuracies can lead to erroneously predicted vaccine candidates by bioinformatics programs. This study evaluates the current N. caninum annotation for potential inaccuracies. Comparisons to annotation from a closely related pathogen, Toxoplasma gondii, are also made to distinguish patterns of inconsistency. More importantly, an RNA-Seq experiment is used to validate the annotation. Potential discrepancies originating from questionable start codon context and exon boundaries were identified in 1943 protein-coding sequences. Four documents are provided here in Dataverse that supplements the paper, ‘Improving the gene-structure annotation of the Apicomplexan parasite Neospora caninum fulfils a vital requirement towards an in silico derived vaccine’.Neospora caninum is an apicomplexan parasite which can cause abortion in cattle instigating major economic burden. Vaccination has been proposed as the most cost-effective control measure to alleviate this burden. Consequently the overriding aspiration for N. caninum research is the identification and subsequent evaluation of vaccine candidates in animal models. To save on time, cost and effort it is now feasible to use an in silico approach for vaccine candidate prediction. Precise protein sequences, derived from the correct open reading frame, are paramount and arguably the most important factor determining success or failure of the approach. The challenge is that publicly available N. caninum sequences are mostly derived from gene predictions. Annotated inaccuracies can lead to erroneously predicted vaccine candidates by bioinformatics programs. This study evaluates the current N. caninum annotation for potential inaccuracies. Comparisons to annotation from a closely related pathogen, Toxoplasma gondii, are also made to distinguish patterns of inconsistency. More importantly, an RNA-Seq experiment is used to validate the annotation. Potential discrepancies originating from questionable start codon context and exon boundaries were identified in 1943 protein-coding sequences. Four documents are provided here in Dataverse that supplements the paper, ‘Improving the gene-structure annotation of the Apicomplexan parasite Neospora caninum fulfils a vital requirement towards an in silico derived vaccine’.