Sacbrood virus (SBV) of honey bees is a picornavirus in the genus Iflavirus. Given its relatively small and simple genome structure, single positive-strand RNA with only one ORF, cloning the full genomic sequence is not difficult. However, adding nonsynonymous mutations to the bee iflavirus clone is difficult because of the lack of information about the viral protein processes. Furthermore, the addition of a reporter gene to the clones has never been accomplished. In preliminary trials, we found that the site between 3′ untranslated region (UTR) and poly(A) can retain added sequences. We added enhanced green fluorescent protein (EGFP) expression at this site, creating a SBV clone with an expression tag that does not affect virus genes. An intergenic region internal ribosome entry site (IRES) from Black queen cell virus (BQCV) was inserted to initiate EGFP expression. The SBV-IRES-EGFP clone successfully infected Apis cerana and Apis mellifera, and in A. cerana larvae, it was isolated and passaged using oral inoculation. The inoculated larvae had higher mortality and the dead larvae showed sacbrood symptoms. The added IRES-EGFP remained in the clone through multiple passages and expressed the expected EGFP in all infected bees. We demonstrated the ability to add gene sequences in the site between 3′-UTR and poly(A) in SBV and the potential to do so in other bee iflaviruses; however, further investigations of the mechanisms are needed. A clone with a desired protein expression reporter will be a valuable tool in bee virus studies.

The 2016 WHO classification of central nervous system tumors has included four molecular subgroups under medulloblastoma (MB) as sonic hedgehog (SHH), wingless (WNT), Grade 3, and Group 4. We aimed to develop machine learning models for predicting MB molecular subgroups based on multi-parameter magnetic resonance imaging (MRI) radiomics, tumor locations, and clinical factors. A total of 122 MB patients were enrolled retrospectively. After selecting robust, non-redundant, and relevant features from 5,529 extracted radiomics features, a random forest model was constructed based on a training cohort ( n = 92) and evaluated on a testing cohort ( n = 30). By combining radiographic features and clinical parameters, two combined prediction models were also built. The subgroup can be classified using an 11-feature radiomics model with a high area under the curve (AUC) of 0.8264 for WNT and modest AUCs of 0.6683, 0.6004, and 0.6979 for SHH, Group 3, and Group 4 in the testing cohort, respectively. Incorporating location and hydrocephalus into the radiomics model resulted in improved AUCs of 0.8403 and 0.8317 for WNT and SHH, respectively. After adding gender and age, the AUCs for WNT and SHH were further improved to 0.9097 and 0.8654, while the accuracies were 70 and 86.67% for Group 3 and Group 4, respectively. Prediction performance was excellent for WNT and SHH, while that for Group 3 and Group 4 needs further improvements. Machine learning algorithms offer potentials to non-invasively predict the molecular subgroups of MB.