BackgroundThis study aimed to compare one state-of-the-art deep learning method and four classical machine learning methods for classifying mediastinal lymph node metastasis of non-small cell lung cancer (NSCLC) from 18F-FDG PET/CT images. Another objective was to compare the discriminative power of the recently popular PET/CT texture features with the widely used diagnostic features such as tumor size, CT value, SUV, image contrast, and intensity standard deviation. The four classical machine learning methods included random forests, support vector machines, adaptive boosting, and artificial neural network. The deep learning method was the convolutional neural networks (CNN). The five methods were evaluated using 1397 lymph nodes collected from PET/CT images of 168 patients, with corresponding pathology analysis results as gold standard. The comparison was conducted using 10 times 10-fold cross-validation based on the criterion of sensitivity, specificity, accuracy (ACC), and area under the ROC curve (AUC). For each classical method, different input features were compared to select the optimal feature set. Based on the optimal feature set, the classical methods were compared with CNN, as well as with human doctors from our institute.ResultsFor the classical methods, the diagnostic features resulted in 81~85% ACC and 0.87~0.92 AUC, which were significantly higher than the results of texture features. CNN’s sensitivity, specificity, ACC, and AUC were 84, 88, 86, and 0.91, respectively. There was no significant difference between the results of CNN and the best classical method. The sensitivity, specificity, and ACC of human doctors were 73, 90, and 82, respectively. All the five machine learning methods had higher sensitivities but lower specificities than human doctors.ConclusionsThe present study shows that the performance of CNN is not significantly different from the best classical methods and human doctors for classifying mediastinal lymph node metastasis of NSCLC from PET/CT images. Because CNN does not need tumor segmentation or feature calculation, it is more convenient and more objective than the classical methods. However, CNN does not make use of the import diagnostic features, which have been proved more discriminative than the texture features for classifying small-sized lymph nodes. Therefore, incorporating the diagnostic features into CNN is a promising direction for future research.

With the development of in-vivo free-space fluorescence molecular imaging and multi-modality imaging for small animals, there is a need for new reconstruction methods for real animal-shape models with a large dataset. In this paper we are reporting a novel hybrid adaptive finite element algorithm for fluorescence tomography reconstruction, based on a linear scheme. Two different inversion strategies (Conjugate Gradient and Landweber iterations) are separately applied to the first mesh level and the succeeding levels. The new algorithm was validated by numerical simulations of a 3-D mouse atlas, based on the latest free-space setup of fluorescence tomography with 360 degrees geometry projections. The reconstructed results suggest that we are able to achieve high computational efficiency and spatial resolution for models with irregular shape and inhomogeneous optical properties.

The temporal and spatial patterning involved in the specification, differentiation, and myelination by oligodendroglia is coordinated in part by the activation and repression of various transcriptional programs. Olig2 is a basic helix-loop-helix transcription factor necessary for oligodendroglial development and expressed continuously throughout the lineage. Despite evidence for the critical role of Olig2 in oligodendroglial specification and differentiation, the function for Olig2 during later stages of oligodendroglial development, namely, the transition into mature oligodendrocytes (OLs) and the formation of the myelin sheath, remains unclear. To address the possibility for a stage-specific role, we deleted Olig2 in oligodendrocyte precursor cells (OPCs) under the control of the CNPase-promoter or in immature OLs under the inducible proteolipid protein promoter. As expected, ablation of Olig2 in OPCs significantly inhibits differentiation, resulting in hypomyelination. However, deletion of the Olig2 gene in immature OLs significantly enhances the maturation process and accelerates the kinetics of myelination/remyelination. Underlying the stage-specific roles for Olig2 is the compensatory expression and function of Olig1, a transcription factor that promotes OL maturation and (re)myelination. Olig1 expression is significantly reduced upon Olig2 deletion in OPCs but is dramatically increased by nearly threefold when deleted in immature OLs. By enforcing expression of Olig1 into OPCs in a null Olig2 background, we demonstrate that overexpression of Olig1 is sufficient to rescue the differentiation phenotype and partially compensates for the Olig2 deletion in vitro. Our results suggest a stage-specific regulatory role for Olig2, mediated by Olig1 that conveys opposing functions on the differentiation and maturation of oligodendrocytes.

Autophagy is a vacuolar/lysosomal cytoplasmic recycling system in eukaryotic cells. ScATG9 is indispensable for autophagy in Saccharomyces cerevisiae. Here, we deleted MgATG9, the orthologue of ScATG9, via targeted gene replacement in the phytopathogenic filamentous fungus Magnaporthe oryzae, and then analyzed the cellular distribution pattern of EGFP-MgAtg9 in the Mgatg9Delta mutant. We detected an expression profile of multiple green dots in the conidial cell inoculated in rich media and in the appressoria differentiated from the conidia in H(2)O. Concurrent with the punctation, we found some fluorescent signals localized on the central vacuole of the submerged hyphae from the conidia cultured in rich media. Next, we introduced DsRed2-MgAtg8 into the Mgatg9Delta mutant expressing EGFP-MgAtg9 and observed partial overlap at multiple sites in the conidial cell, reminiscent of that in the mammalian system. Our findings further led to the postulation that the multiple sites where the two fusions colocalized tend to merge as a central structure in the conidial cell. Finally, we tested the expression of EGFP-MgAtg9 in null mutants of MgATG1, 2, 13 and 18, respectively. We speculate that MgAtg1, 2 and 18, but not MgAtg13, is required for MgAtg9 cycling through the multiple colocalization sites to its storage pools in the conidial cell of M. oryzae, and fusion of these colocalization sites into a central structure could be governed through other unidentified mechanisms.

No abstract

No abstract

No abstract

Partial 28S rRNA gene sequence-data of the strains of the anamorphic genera Bahusutrabeeja, Diplococcium, Natarajania, Paliphora, Polyschema, Rattania and Spadicoides were analysed to predict their phylogenetic relationships and taxonomic placement within the Ascomycota. Results indicate that Diplococcium and morphologically similar genera, i.e. Spadicoides, Paliphora and Polyschema do not share a recent common ancestor. The type species of Diplococcium, D. spicatum is referred to Helotiales (Leotiomycetes). The placement of Spadicoides bina, the type of the genus, is unresolved but it is shown to be closely associated with Porosphaerella species, which are sister taxa to Coniochaetales (Sordariomycetes). Three Polyschema species analysed in this study represent a monophyletic lineage and are related to Lentithecium fluviatile and Leptosphaeria calvescens in Pleosporales (Dothideomycetes). DNA sequence analysis also suggests that Paliphora intermedia is a member of Chaetosphaeriaceae (Sordariomycetes). The type species of Bahusutrabeeja, B. dwaya, is phylogenetically related to Neodeightonia (=Botryosphaeria) subglobosa in Botryosphaeriales (Dothideomycetes). Monotypic genera Natarajania and Rattania are phylogenetically related to members of Diaporthales and Chaetosphaeriales, respectively. Future studies with extended gene datasets and type strains are required to resolve many novel but morphologically unexplainable phylogenetic scenarios revealed from this study. It is increasingly becoming evident that a fungal lineage may include a mosaic of anamorphs, teleomorphs and pleomorphs whose morphologies may not always be correlated. It is therefore suggested that where possible all new species descriptions, whether teleomorphic, anamorphic or pleomorphic, should include DNA sequence-data to facilitate amalgamation of anamorphic and pleomorphic genera in a single phylogenetic classification system.