Search

Abstract Background Necroptosis is a novel type of programmed cell death distinct from apoptosis. However, the role of necroptosis in ovarian cancer (OC) remains unclear. The present study investigated the prognostic value of necroptosis-related genes (NRGs) and the immune landscape in OC. Methods The gene expression profiling and clinical information were downloaded from the TCGA and GTEx databases. Differentially expressed NRGs (DE-NRGs) between OC and normal tissueswere identified. The regression analyses were conducted to screen the prognostic NRGs and construct the predictive risk model. Patients were then divided into high- and low-risk groups, and the GO and KEGG analyses were performed to explore bioinformatics functions between the two groups. Subsequently, the risk level and immune status correlations were assessed through the ESTIMATE and CIBERSORT algorithms. The tumor mutation burden (TMB) and the drug sensitivity were also analyzed based on the two-NRG signature in OC. Results Totally 42 DE-NRGs were identified in OC. The regression analyses screened out two NRGs (MAPK10 and STAT4) with prognostic values for overall survival. The ROC curve showed a better predictive ability in five-year OS using the risk score. Immune-related functions were significantly enriched in the high- and low-risk group. Macrophages M1, T cells CD4 memory activated, T cells CD8, and T cells regulatory infiltration immune cells were associated with the low-risk score. The lower tumor microenvironment score was demonstrated in the high-risk group. Patients with lower TMB in the low-risk group showed a better prognosis, and a lower TIDE score suggested a better immune checkpoint inhibitor response in the high-risk group. Besides, cisplatin and paclitaxel were found to be more sensitive in the low-risk group. Conclusions MAPK10 and STAT4 can be important prognosis factors in OC, and the two-gene signature performs well in predicting survival outcomes. Our study provided novel ways of OC prognosis estimation and potential treatment strategy.

In this paper, we investigate a challenging but interesting task in the research of speech emotion recognition (SER), i.e., cross-corpus SER. Unlike the conventional SER, the training (source) and testing (target) samples in cross-corpus SER come from different speech corpora, which results in a feature distribution mismatch between them. Hence, the performance of most existing SER methods may sharply decrease. To cope with this problem, we propose a simple yet effective deep transfer learning method called progressive distribution adapted neural networks (PDAN). PDAN employs convolutional neural networks (CNN) as the backbone and the speech spectrum as the inputs to achieve an end-to-end learning framework. More importantly, its basic idea for solving cross-corpus SER is very straightforward, i.e., enhancing the backbone's corpus invariant feature learning ability by incorporating a progressive distribution adapted regularization term into the original loss function to guide the network training. To evaluate the proposed PDAN, extensive cross-corpus SER experiments on speech emotion corpora including EmoDB, eNTERFACE, and CASIA are conducted. Experimental results showed that the proposed PDAN outperforms most well-performing deep and subspace transfer learning methods in dealing with the cross-corpus SER tasks.