Clinically Applicable Pathological Diagnosis System for Cell Clumps in Endometrial Cancer Screening via Deep Convolutional Neural Networks.

OBJECTIVES: The soaring demand for endometrial cancer screening has exposed a huge shortage of cytopathologists worldwide. To address this problem, our study set out to establish an artificial intelligence system that automatically recognizes and diagnoses pathological images of endometrial cell clumps (ECCs). METHODS: We used Li Brush to acquire endometrial cells from patients. Liquid-based cytology technology was used to provide slides. The slides were scanned and divided into malignant and benign groups. We proposed two (a U-net segmentation and a DenseNet classification) networks to identify images. Another four classification networks were used for comparison tests. RESULTS: A total of 113 (42 malignant and 71 benign) endometrial samples were collected, and a dataset containing 15,913 images was constructed. A total of 39,000 ECCs patches were obtained by the segmentation network. Then, 26,880 and 11,520 patches were used for training and testing, respectively. On the premise that the training set reached 100%, the testing set gained 93.5% accuracy, 92.2% specificity, and 92.0% sensitivity. The remaining 600 malignant patches were used for verification. CONCLUSIONS: An artificial intelligence system was successfully built to classify malignant and benign ECCs.

Combining Radiology and Pathology for Automatic Glioma Classification.

Subtype classification is critical in the treatment of gliomas because different subtypes lead to different treatment options and postoperative care. Although many radiological- or histological-based glioma classification algorithms have been developed, most of them focus on single-modality data. In this paper, we propose an innovative two-stage model to classify gliomas into three subtypes (i.e., glioblastoma, oligodendroglioma, and astrocytoma) based on radiology and histology data. In the first stage, our model classifies each image as having glioblastoma or not. Based on the obtained non-glioblastoma images, the second stage aims to accurately distinguish astrocytoma and oligodendroglioma. The radiological images and histological images pass through the two-stage design with 3D and 2D models, respectively. Then, an ensemble classification network is designed to automatically integrate the features of the two modalities. We have verified our method by participating in the MICCAI 2020 CPM-RadPath Challenge and won 1st place. Our proposed model achieves high performance on the validation set with a balanced accuracy of 0.889, Cohen's Kappa of 0.903, and an F1-score of 0.943. Our model could advance multimodal-based glioma research and provide assistance to pathologists and neurologists in diagnosing glioma subtypes. The code has been publicly available online at https://github.com/Xiyue-Wang/1st-in-MICCAI2020-CPM.

3D Dense Volumetric Network for Accurate Automated Pancreas Segmentation.

Pancreatic cancer poses a great threat to our health with an overall five-year survival rate of 8%. Automatic and accurate segmentation of pancreas plays an important and prerequisite role in computer-assisted diagnosis and treatment. Due to the ambiguous pancreas borders and intertwined surrounding tissues, it is a challenging task. In this paper, we propose a novel 3D Dense Volumetric Network (3D2VNet) to improve the segmentation accuracy of pancreas organ. Firstly, 3D fully convolutional architecture is applied to effectively incorporate the 3D pancreas and geometric cues for volume-to-volume segmentation. Then, dense connectivity is introduced to preserve the maximum information flow between layers and reduce the overfitting on limited training data. In addition, a auxiliary side path is constructed to help the gradient propagation to stabilize the training process. Adequate experiments are conducted on a challenging pancreas dataset in Medical Segmentation Decathlon challenge. The results demonstrate our method can outperform other comparison methods on the task of automated pancreas segmentation using limited data.Clinical relevance-This paper proposes an accurate automated pancreas segmentation method, which can provide assistance to clinicians in the diagnosis and treatment of pancreatic cancer.

A computational method for prediction of matrix proteins in endogenous retroviruses.

Human endogenous retroviruses (HERVs) encode active retroviral proteins, which may be involved in the progression of cancer and other diseases. Matrix protein (MA), in group-specific antigen genes (gag) of retroviruses, is associated with the virus envelope glycoproteins in most mammalian retroviruses and may be involved in virus particle assembly, transport and budding. However, the amount of annotated MAs in ERVs is still at a low level so far. No computational method to predict the exact start and end coordinates of MAs in gags has been proposed yet. In this paper, a computational method to identify MAs in ERVs is proposed. A divide and conquer technique was designed and applied to the conventional prediction model to acquire better results when dealing with gene sequences with various lengths. Initiation sites and termination sites were predicted separately and then combined according to their intervals. Three different algorithms were applied and compared: weighted support vector machine (WSVM), weighted extreme learning machine (WELM) and random forest (RF). G - mean (geometric mean of sensitivity and specificity) values of initiation sites and termination sites under 5-fold cross validation generated by random forest models are 0.9869 and 0.9755 respectively, highest among the algorithms applied. Our prediction models combine RF & WSVM algorithms to achieve the best prediction results. 98.4% of all the collected ERV sequences with complete MAs (125 in total) could be predicted exactly correct by the models. 94,671 HERV sequences from 118 families were scanned by the model, 104 new putative MAs were predicted in human chromosomes. Distributions of the putative MAs and optimizations of model parameters were also analyzed. The usage of our predicting method was also expanded to other retroviruses and satisfying results were acquired.

A computational model for predicting integrase catalytic domain of retrovirus.

Integrase catalytic domain (ICD) is an essential part in the retrovirus for integration reaction, which enables its newly synthesized DNA to be incorporated into the DNA of infected cells. Owing to the crucial role of ICD for the retroviral replication and the absence of an equivalent of integrase in host cells, it is comprehensible that ICD is a promising drug target for therapeutic intervention. However, annotated ICDs in UniProtKB database have still been insufficient for a good understanding of their statistical characteristics so far. Accordingly, it is of great importance to put forward a computational ICD model in this work to annotate these domains in the retroviruses. The proposed model then discovered 11,660 new putative ICDs after scanning sequences without ICD annotations. Subsequently in order to provide much confidence in ICD prediction, it was tested under different cross-validation methods, compared with other database search tools, and verified on independent datasets. Furthermore, an evolutionary analysis performed on the annotated ICDs of retroviruses revealed a tight connection between ICD and retroviral classification. All the datasets involved in this paper and the application software tool of this model can be available for free download at https://sourceforge.net/projects/icdtool/files/?source=navbar.

Genome-wide identification and predictive modeling of lincRNAs polyadenylation in cancer genome.

Long noncoding RNAs (lncRNAs) play essential regulatory roles in the human cancer genome. Many identified lncRNAs are transcribed by RNA polymerase II in which they are polyadenylated, whereby the long intervening noncoding RNAs (lincRNAs) have been widely used for the researches of lncRNAs. To date, the mechanism of lincRNAs polyadenylation related to cancer is rarely fully understood yet. In this paper, first we reported a comprehensive map of global lincRNAs polyadenylation sites (PASs) in five human cancer genomes; second we proposed a grouping method based on the pattern of genes expression and the manner of alternative polyadenylation (APA); third we investigated the distribution of motifs surrounding PASs. Our analysis reveals that about 70% of PASs are located in the sense strand of lincRNAs. Also more than 90% PASs in the antisense strand of lincRNAs are located in the intron regions. In addition, around 40% of lincRNA genes with PASs has APA sites. Four obvious motifs i.e., AATAAA, TTTTTTTT, CCAGSCTGG, and RGYRYRGTGG were detected in the sequences surrounding PASs in the normal and cancer tissues. Furthermore, a novel algorithm was proposed to recognize the lincRNAs PASs of tumor tissues based on support vector machine (SVM). The algorithm can achieve the accuracies up to 96.55% and 89.48% for identification the tumor lincRNAs PASs from the non-polyadenylation sites and the non-lincRNA PASs, respectively.

ISDTool 2.0: a computational model for predicting immunosuppressive domain of retroviruses.

Immunosuppressive domain (ISD) is a conserved region of transmembrane proteins (TM) in envelope gene (env) of retroviruses. in vitro and vivo, a synthetic peptide (CKS-17) that shows homology to ISD inhibits immune function. Evidence has shown that ISD suppresses lymphocyte proliferation and allows escape from immune effectors of the innate and adaptive arms in mouse immune system. Previously, we have developed a tool ISDTool 1.0 to identify ISD of human endogenous retrovirus (HERV). However, several other important retroviruses exist and no method is devoted to ISD prediction of them so far. In the paper, a computational model is proposed to identify ISD of six typical retroviruses from three species. The model combines the minimum Redundancy Maximum Relevance (mRMR) feature selection criterion with weighted extreme learning machine (WELM) to achieve high identification accuracies of 98.95%, 96.34% and 96.87% using self-consistency, 5-fold and 10-fold cross-validation, respectively. A software tool named ISDTool 2.0 has been developed to facilitate the application of the model and a large number of new putative ISDs of the six retroviruses were predicted. In addition, motifs of ISD in these retroviruses were analyzed and the evolutionary relationship was discussed. Datasets and the software involved in the paper are available at http://sourceforge.net/projects/isdtool/files/ISDTool-2.0/.

ISDTool: a computational model for predicting immunosuppressive domain of HERVs.

Human endogenous retroviruses (HERVs) have been found to act as etiological cofactors in several chronic diseases, including cancer, autoimmunity and neurological dysfunction. Immunosuppressive domain (ISD) is a conserved region of transmembrane protein (TM) in envelope gene (env) of retroviruses. In vitro and vivo, evidence has shown that retroviral TM is highly immunosuppressive and a synthetic peptide (CKS-17) that shows homology to ISD inhibits immune function. ISD is probably a potential pathogenic element in HERVs. However, only less than one hundred ISDs of HERVs have been annotated by researchers so far, and universal software for domain prediction could not achieve sufficient accuracy for specific ISD. In this paper, a computational model is proposed to identify ISD in HERVs based on genome sequences only. It has a classification accuracy of 97.9% using Jack-knife test. 117 HERVs families were scanned with the model, 1002 new putative ISDs have been predicted and annotated in the human chromosomes. This model is also applicable to search for ISDs in human T-lymphotropic virus (HTLV), simian T-lymphotropic virus (STLV) and murine leukemia virus (MLV) because of the evolutionary relationship between endogenous and exogenous retroviruses. Furthermore, software named ISDTool has been developed to facilitate the application of the model. Datasets and the software involved in the paper are all available at https://sourceforge.net/projects/isdtool/files/ISDTool-1.0.

A hybrid model for the prediction of mRNA polyadenylation signals.

The mRNA polyadenylation is the cellular process that adds adenosine tails to mature mRNAs. Malfunction of polyadenylation has been implicated in several human diseases. In this paper, we proposed a novel feature extraction approach which employs the K-gram nucleotide acid pattern, the position weight matrix (PWM) and the increment of diversity (ID) to represent the original features. Then Principle Component Analysis (PCA) was applied to transform the original features into a new feature space where the low-dimensional features were used to train the real-coded genetic neural network model. In the experiments, our proposed algorithm (GA-BP) can achieve the accuracy about 82.98%, specificity 82.95% and sensitivity 83.01% in the specific dataset constructed by Kalkatawi. The results demonstrate that GA-BP is a promising algorithm for the prediction of mRNA polyadenylation signals.