Biology subjects

Sheng, K.

Publications and source records attributed to Sheng, K..

4 recordsLinked to original sources

bioRxiv2021

U-RISC: An Ultra-high Resolution EM Dataset Challenging Existing Deep Learning Algorithms

Connectomics is a developing field aiming at reconstructing the connection of the neural system at nanometer scale. Computer vision technology, especially deep learning methods used in image processing, has promoted connectomic data analysis to a new era. However, the performance of the state-of-the-art methods still falls behind the demand of scientific research. Inspired by the success of ImageNet, we present the U-RISC, an annotated Ultra-high Resolution Image Segmentation dataset for Cell membrane, which is the largest cell membrane annotated Electron Microscopy (EM) dataset with a resolution of 2.18nm/pixel. Multiple iterative annotations ensured the quality of the dataset. Through an open competition, we reveal that the performance of current deep learning methods still has a considerable gap with human-level, different from ISBI 2012, on which the performance of deep learning is close to human. To explore the causes of this discrepancy, we analyze the neural networks with a visualization method, attribution analysis. We find that in U-RISC, it requires a larger area around a pixel to predict whether the pixel belongs to the cell membrane or not. Finally, we integrate currently available methods to provide a new benchmark (0.67, 10% higher than the leader of competition, 0.61) for cell membrane segmentation on U-RISC and propose some suggestions in developing deep learning algorithms. The U-RISC dataset and the deep learning codes used in this paper will be publicly available.

neuroscience↗

bioRxiv2021

A General LSTM-based Deep Learning Method for Estimating Neuronal Models and Inferring Neural Circuitry

Computational neural models are essential tools for neuroscientists to study the functional roles of single neurons or neural circuits. With the recent advances in experimental techniques, there is a growing demand to build up neural models at single neuron or large-scale circuit levels. A long-standing challenge to build up such models lies in tuning the free parameters of the models to closely reproduce experimental recordings. There are many advanced machine-learning-based methods developed recently for parameter tuning, but many of them are task-specific or requires onerous manual interference. There lacks a general and fully-automated method since now. Here, we present a Long Short-Term Memory (LSTM)-based deep learning method, General Neural Estimator (GNE), to fully automate the parameter tuning procedure, which can be directly applied to both single neuronal models and large-scale neural circuits. We made comprehensive comparisons with many advanced methods, and GNE showed outstanding performance on both synthesized data and experimental data. Finally, we proposed a roadmap centered on GNE to help guide neuroscientists to computationally reconstruct single neurons and neural circuits, which might inspire future brain reconstruction techniques and corresponding experimental design. The code of our work will be publicly available upon acceptance of this paper.

neuroscience↗

bioRxiv2020

Connexin 43 confers chemoresistance through activating PI3K

Circumventing chemoresistance is crucial for effectively treating glioblastoma due to limited therapeutic options. The gap junction protein connexin 43 (Cx43) renders glioblastoma resistant to the frontline chemotherapy temozolomide; however, targeting Cx43 is difficult because mechanisms underlying Cx43-mediated chemoresistance remain elusive. Here we show that Cx43, but not other connexins, is highly expressed in glioblastoma and strongly correlates with poor patient prognosis and chemoresistance, making Cx43 the prime therapeutic target among all connexins. The intracellular carboxyl terminus of Cx43 binds to phosphatidylinositol 3-kinase (PI3K) catalytic subunit {beta} (PIK3CB, also called PI3K{beta} or p110{beta}), thereby activating PI3K signaling independent of Cx43-channels and subsequently inducing temozolomide resistance. A combination of CT1, a Cx43-targeting peptide inhibitor, and PIK3CB-selective inhibitors restores temozolomide sensitivity in vitro and in vivo. This study not only reveals novel mechanistic insights into chemoresistance in glioblastoma, but also demonstrates that targeting Cx43 and PIK3CB/p110{beta} is an effective approach for overcoming chemoresistance.

cancer biology↗

bioRxiv2019

Topological Dependent Abundance of Spontaneous DNA Damage in Single Human Cells

In the studies of single-cell genomics, the large endeavor has been focused on the detection of the permanent changes in the genome. On the other hand, spontaneous DNA damage frequently occurs and results in transient single-stranded changes to the genome until they are repaired. So far, successful profiling of these dynamic changes has not been demonstrated by single-cell whole-genome amplification methods. Here we reported a novel single-cell WGA method: Linearly Produced Semiamplicon based Split Amplification Reaction (LPSSAR), which allows, for the first time, the genome-wide detection of the DNA damage associated single nucleotide variants (dSNVs) in single human cells. The sequence-based detection of dSNVs allows the direct characterization of the major damage signature that occurred in human cells. In the analysis of the abundance of dSNVs along the genome, we observed two modules of dSNV abundance, instead of a homogeneous abundance of dSNVs. Interestingly, we found that the two modules are associated with the A/B topological compartments of the genome. This result suggests that the genome topology directly influences genome stability. Furthermore, with the detection of a large number of dSNVs in single cells, we showed that only under a stringent filtering condition, can we distinguish the de novo mutations from the dSNVs and achieve a reliable estimation of the total level of de novo mutations in a single cell.

genomics↗