English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Benefiting from the advancements in deep learning, various genomic analyticaltechniques, such as survival analysis, classification of tumors and theirsubtypes, and exploration of specific pathways, have significantly enhanced ourunderstanding of the biological mechanisms driving cancer. However, theoverfitting issue, arising from the limited number of patient samples, poses achallenge in improving the accuracy of genome analysis by deepening the neuralnetwork. Furthermore, it remains uncertain whether novel approaches such as thesparsely gated mixture of expert (MOE) and self-attention mechanisms canimprove the accuracy of genomic analysis. In this paper, we introduce a novelsparsely gated RNA-seq analysis framework called Gene-MOE. This frameworkexploits the potential of the MOE layers and the proposed mixture of attentionexpert (MOAE) layers to enhance the analysis accuracy. Additionally, itaddresses overfitting challenges by integrating pan-cancer information from 33distinct cancer types through pre-training.We pre-trained Gene-MOE on TCGApan-cancer RNA-seq dataset with 33 cancer types. Subsequently, we conductedexperiments involving cancer classification and survival analysis based on thepre-trained Gene-MOE. According to the survival analysis results on 14 cancertypes, Gene-MOE outperformed state-of-the-art models on 12 cancer types.Through detailed feature analysis, we found that the Gene-MOE model could learnrich feature representations of high-dimensional genes. According to theclassification results, the total accuracy of the classification model for 33cancer classifications reached 95.8%, representing the best performancecompared to state-of-the-art models. These results indicate that Gene-MOE holdsstrong potential for use in cancer classification and survival analysis.

Gene-MOE: A sparsely gated prognosis and classification framework exploiting pan-cancer genomic information