Biology subjects

Shen, G.

Publications and source records attributed to Shen, G..

4 recordsLinked to original sources

bioRxiv2018

Deciphering the metabolic perturbation in hepatic alveolar echinococcosis: a 1H NMR-based metabolomics study

Hepatic alveolar echinococcosis (HAE) is a chronic and potentially lethal parasitic disease. It is caused by growth of Echinococcus multilocularis larvae in liver. To date, early-stage diagnosis for the disease is not mature due to its long asymptomatic incubation period. In this study, a proton nuclear magnetic resonance (1H NMR) -based metabolomics approach was applied in conjunction with multivariate statistical analysis to investigate the altered metabolic profiles in blood serum and urine samples from HAE patients and to identify characteristic metabolic markers associated with HAE. The current results identified 21 distinctive metabolic difference between the HAE patients and healthy individuals, which can be associated with perturbations in energy metabolism, amino acid metabolism, oxidative stress, and neurotransmitter imbalance. In addition, the Fischer ratio, which is the molar ratio of branched-chain amino acids to aromatic amino acids was found significantly lower (p<0.001) in blood serum from HAE patients. The ratio, together with changes in other metabolic pathways may provide new insight into mechanistic understanding of HAE pathogenesis, and may be useful for early-stage HAE diagnosis.\n\nAuthor SummaryHepatic alveolar echinococcosis (HAE) is a life-threatening disease caused by Echinococcus multilocularis infection. The disease has a long asymptomatic early stage (5~15 years), which complicates effective diagnosis of early-stage HAE even with advanced imaging techniques. Metabolomics is an emerging analytical platform that comprises of analysis of all small molecule metabolites that are present within an organism. The applications of metabolomics method on HAE may help to reveal the molecular biology mechanisms of HAE. In the current study, we had used 1H NMR-based metabolomics technique to investigate blood serum and urine samples from HAE patients. Altered metabolic responses and characteristic differential metabolites for HAE were identified. The metabolic profiling of human biofluids provided valuable information for early-stage HAE diagnosis and for therapeutic interventions, without having to extract HAE vesicles from patients. By featuring global and comprehensive metabolic status, the metabolomics approach holds considerable promise as a noninvasive, dynamic, and effective tool for probing the underlying mechanism of HAE.

systems biology↗

bioRxiv2018

Large-scale gene losses underlie the genome evolution of parasitic plant Cuscuta australis

Dodders (Cuscuta spp., Convolvulaceae) are globally distributed root- and leafless parasitic plants that parasitize a wide range of hosts. The physiology, ecology, and evolution of these obligate parasites are still poorly understood. A high-quality reference genome (size 266.74 Mb and contig N50 of 3.63 Mb) of Cuscuta australis was assembled. Our analyses reveal that Cuscuta experienced accelerated evolution, and Cuscuta and the convolvulaceous morning glory (Ipomoea) shared a common whole-genome triplication event before their divergence. Importantly, C. australis genome harbors only 19805 protein-coding genes, and 11.7% of the conserved orthologs in autotrophic plants are lost in C. australis. Many of these gene loss events likely result from the plants parasitic lifestyle and large changes in its body plan. Moreover, comparison of the gene expression patterns in Cuscuta prehaustoria/haustoria and various tissues of closely related autotrophic plants suggests that Cuscuta haustorium genes largely evolved from roots. The C. australis genome provides important resources for studying the evolution of parasitism, regressive evolution, and evo-devo in plant parasites.

genomics↗

bioRxiv2018

End-to-end deep image reconstruction from human brain activity

Deep neural networks (DNNs) have recently been applied successfully to brain decoding and image reconstruction from functional magnetic resonance imaging (fMRI) activity. However, direct training of a DNN with fMRI data is often avoided because the size of available data is thought to be insufficient to train a complex network with numerous parameters. Instead, a pre-trained DNN has served as a proxy for hierarchical visual representations, and fMRI data were used to decode individual DNN features of a stimulus image using a simple linear model, which were then passed to a reconstruction module. Here, we present our attempt to directly train a DNN model with fMRI data and the corresponding stimulus images to build an end-to-end reconstruction model. We trained a generative adversarial network with an additional loss term defined in a high-level feature space (feature loss) using up to 6,000 training data points (natural images and the fMRI responses). The trained deep generator network was tested on an independent dataset, directly producing a reconstructed image given an fMRI pattern as the input. The reconstructions obtained from the proposed method showed resemblance with both natural and artificial test stimuli. The accuracy increased as a function of the training data size, though not outperforming the decoded feature-based method with the available data size. Ablation analyses indicated that the feature loss played a critical role to achieve accurate reconstruction. Our results suggest a potential for the end-to-end framework to learn a direct mapping between brain activity and perception given even larger datasets.

neuroscience↗

bioRxiv2017

Deep image reconstruction from human brain activity

Machine learning-based analysis of human functional magnetic resonance imaging (fMRI) patterns has enabled the visualization of perceptual content. However, it has been limited to the reconstruction with low-level image bases (Miyawaki et al., 2008; Wen et al., 2016) or to the matching to exemplars (Naselaris et al., 2009; Nishimoto et al., 2011). Recent work showed that visual cortical activity can be decoded (translated) into hierarchical features of a deep neural network (DNN) for the same input image, providing a way to make use of the information from hierarchical visual features (Horikawa & Kamitani, 2017). Here, we present a novel image reconstruction method, in which the pixel values of an image are optimized to make its DNN features similar to those decoded from human brain activity at multiple layers. We found that the generated images resembled the stimulus images (both natural images and artificial shapes) and the subjective visual content during imagery. While our model was solely trained with natural images, our method successfully generalized the reconstruction to artificial shapes, indicating that our model indeed reconstructs or generates images from brain activity, not simply matches to exemplars. A natural image prior introduced by another deep neural network effectively rendered semantically meaningful details to reconstructions by constraining reconstructed images to be similar to natural images. Furthermore, human judgment of reconstructions suggests the effectiveness of combining multiple DNN layers to enhance visual quality of generated images. The results suggest that hierarchical visual information in the brain can be effectively combined to reconstruct perceptual and subjective images.

neuroscience↗