Leveraging Attention-Based Convolutional Neural Networks for Meningioma Classification in Computational Histopathology.

Convolutional neural networks (CNNs) are becoming increasingly valuable tools for advanced computational histopathology, promoting precision medicine through exceptional visual decoding abilities. Meningiomas, the most prevalent primary intracranial tumors, necessitate accurate grading and classification for informed clinical decision-making. Recently, DNA methylation-based molecular classification of meningiomas has proven to be more effective in predicting tumor recurrence than traditional histopathological methods. However, DNA methylation profiling is expensive, labor-intensive, and not widely accessible. Consequently, a digital histology-based prediction of DNA methylation classes would be advantageous, complementing molecular classification. In this study, we developed and rigorously assessed an attention-based multiple-instance deep neural network for predicting meningioma methylation classes using tumor methylome data from 142 (+51) patients and corresponding hematoxylin-eosin-stained histological sections. Pairwise analysis of sample cohorts from three meningioma methylation classes demonstrated high accuracy in two combinations. The performance of our approach was validated using an independent set of 51 meningioma patient samples. Importantly, attention map visualization revealed that the algorithm primarily focuses on tumor regions deemed significant by neuropathologists, offering insights into the decision-making process of the CNN. Our findings highlight the capacity of CNNs to effectively harness phenotypic information from histological sections through computerized images for precision medicine. Notably, this study is the first demonstration of predicting clinically relevant DNA methylome information using computer vision applied to standard histopathology. The introduced AI framework holds great potential in supporting, augmenting, and expediting meningioma classification in the future.

DistSNE: Distributed computing and online visualization of DNA methylation-based central nervous system tumor classification.

The current state-of-the-art analysis of central nervous system (CNS) tumors through DNA methylation profiling relies on the tumor classifier developed by Capper and colleagues, which centrally harnesses DNA methylation data provided by users. Here, we present a distributed-computing-based approach for CNS tumor classification that achieves a comparable performance to centralized systems while safeguarding privacy. We utilize the t-distributed neighborhood embedding (t-SNE) model for dimensionality reduction and visualization of tumor classification results in two-dimensional graphs in a distributed approach across multiple sites (DistSNE). DistSNE provides an intuitive web interface (https://gin-tsne.med.uni-giessen.de) for user-friendly local data management and federated methylome-based tumor classification calculations for multiple collaborators in a DataSHIELD environment. The freely accessible web interface supports convenient data upload, result review, and summary report generation. Importantly, increasing sample size as achieved through distributed access to additional datasets allows DistSNE to improve cluster analysis and enhance predictive power. Collectively, DistSNE enables a simple and fast classification of CNS tumors using large-scale methylation data from distributed sources, while maintaining the privacy and allowing easy and flexible network expansion to other institutes. This approach holds great potential for advancing human brain tumor classification and fostering collaborative precision medicine in neuro-oncology.

SangeR: the high-throughput Sanger sequencing analysis pipeline.

Summary: In the era of next generation sequencing and beyond, the Sanger technique is still widely used for variant verification of inconclusive or ambiguous high-throughput sequencing results or as a low-cost molecular genetical analysis tool for single targets in many fields of study. Many analysis steps need time-consuming manual intervention. Therefore, we present here a pipeline-capable high-throughput solution with an optional Shiny web interface, that provides a binary mutation decision of hotspots together with plotted chromatograms including annotations via flat files. Availability and implementation: SangeR is freely available at https://github.com/Neuropathology-Giessen/SangeR and https://hub.docker.com/repository/docker/kaischmid/sange_r. Contact: Kai.Schmid@patho.med.uni-giessen.de or Daniel.Amsel@patho.med.uni-giessen.de. Supplementary information: Supplementary data are available at Bioinformatics online.

Single cell sequencing reveals endothelial plasticity with transient mesenchymal activation after myocardial infarction.

Endothelial cells play a critical role in the adaptation of tissues to injury. Tissue ischemia induced by infarction leads to profound changes in endothelial cell functions and can induce transition to a mesenchymal state. Here we explore the kinetics and individual cellular responses of endothelial cells after myocardial infarction by using single cell RNA sequencing. This study demonstrates a time dependent switch in endothelial cell proliferation and inflammation associated with transient changes in metabolic gene signatures. Trajectory analysis reveals that the majority of endothelial cells 3 to 7 days after myocardial infarction acquire a transient state, characterized by mesenchymal gene expression, which returns to baseline 14 days after injury. Lineage tracing, using the Cdh5-CreERT2;mT/mG mice followed by single cell RNA sequencing, confirms the transient mesenchymal transition and reveals additional hypoxic and inflammatory signatures of endothelial cells during early and late states after injury. These data suggest that endothelial cells undergo a transient mes-enchymal activation concomitant with a metabolic adaptation within the first days after myocardial infarction but do not acquire a long-term mesenchymal fate. This mesenchymal activation may facilitate endothelial cell migration and clonal expansion to regenerate the vascular network.

Noncompaction of the ventricular myocardium is associated with a de novo mutation in the beta-myosin heavy chain gene.

Noncompaction of the ventricular myocardium (NVM) is the morphological hallmark of a rare familial or sporadic unclassified heart disease of heterogeneous origin. NVM results presumably from a congenital developmental error and has been traced back to single point mutations in various genes. The objective of this study was to determine the underlying genetic defect in a large German family suffering from NVM. Twenty four family members were clinically assessed using advanced imaging techniques. For molecular characterization, a genome-wide linkage analysis was undertaken and the disease locus was mapped to chromosome 14ptel-14q12. Subsequently, two genes of the disease interval, MYH6 and MYH7 (encoding the alpha- and beta-myosin heavy chain, respectively) were sequenced, leading to the identification of a previously unknown de novo missense mutation, c.842G>C, in the gene MYH7. The mutation affects a highly conserved amino acid in the myosin subfragment-1 (R281T). In silico simulations suggest that the mutation R281T prevents the formation of a salt bridge between residues R281 and D325, thereby destabilizing the myosin head. The mutation was exclusively present in morphologically affected family members. A few members of the family displayed NVM in combination with other heart defects, such as dislocation of the tricuspid valve (Ebstein's anomaly, EA) and atrial septal defect (ASD). A high degree of clinical variability was observed, ranging from the absence of symptoms in childhood to cardiac death in the third decade of life. The data presented in this report provide first evidence that a mutation in a sarcomeric protein can cause noncompaction of the ventricular myocardium.

Novel deletions in MYH7 and MYBPC3 identified in Indian families with familial hypertrophic cardiomyopathy.

Mutations causing familial hypertrophic cardiomyopathy (HCM) have been described in at least 11 genes encoding cardiac sarcomeric proteins. In this study, three previously unknown deletions have been identified in the human cardiac genes coding for beta-myosin heavy chain (MYH7 on chromosome 14) and myosin-binding protein-C (MYBPC3 on chromosome 11). In family MM, a 3-bp deletion in MYH7 was detected to be associated with loss of glutamic acid in position 927 (DeltaE927) of the myosin rod. In two other families (HH and NP, related by a common founder) a 2-bp loss in codon 453 (exon 16) of MYBPC3 was identified as the presumable cause of a translation reading frame shift. Taken together 15 living mutation carriers were investigated. Six deceased family members (with five cases of premature sudden cardiac death (SCD) in families MM and NP) were either obligate or suspected mutation carriers. In addition to these mutations a 25-bp deletion in intron 32 of MYBPC3 was identified in family MM (five carriers) and in a fourth family (MiR, one HCM patient, three deletion carriers). In agreement with the loss of the regular splicing branch point in the altered intron 32, a splicing deficiency was observed in an exon trapping experiment using MYBPC3 exon 33 as a test substrate. Varying disease profiles assessed using standard clinical, ECG and echocardiographic procedures in conjunction with mutation analysis led to the following conclusions: (1) In family MM the DeltaE927 deletion in MYH7 was assumed to be associated with complete penetrance. Two cases of reported SCD might have been related to this mutation. (2) The two families, HH and NP, distantly related by a common founder, and both suffering from a 2-bp deletion in exon 16 of MYBPC3 differed in their average phenotypes. In family NP, four cases of cardiac death were documented, whereas no cardiac-related death was reported from family HH. These results support the notion that mutations in HCM genes may directly determine disease penetrance and severity; however, a contribution of additional, unidentified factors (genes) to the HCM phenotype can-at least in some cases-not be excluded. (3) The deletion in intron 32 of MYBPC3 was seen in two families, but in both its relation to disease was not unequivocal. In addition, this deletion was observed in 16 of 229 unrelated healthy individuals of the population of the South Indian states of Kerala and Tamil Nadu. It was not seen in 270 Caucasians from Russia and western Europe. Hence, it is considered to represent a regional genetic polymorphism restricted to southern India. The association of the deletion with altered splicing in transfected cells suggests that this deletion may create a "modifying gene", which is per se not or only rarely causing HCM, but which may enhance the phenotype of a mutation responsible for disease.