Biology subjects

Shi, J.

Publications and source records attributed to Shi, J..

28 records · Page 2Linked to original sources

bioRxiv2017

Cryo-ET reveals nucleosome reorganization in condensed mitotic chromosomes in vivo

Chromosomes condense during mitosis in most eukaryotes. This transformation involves rearrangements at the nucleosome level and has consequences for transcription, but the details remain unclear. Here, we use cryo-electron tomography to determine the 3-D arrangement of nucleosomes and other large nuclear features in frozen-hydrated fission-yeast cells. Nucleosomes can form irregular clusters in both interphase and mitotic cells, but they are smaller than expected for Hi-C domains. The nucleosomes are co-mingled with two features: nucleosome-free pockets and megadalton-sized \"megacomplexes\". Compared to interphase, the nucleosomes in mitotic chromosomes pack into slightly larger clusters. However, nearest-neighbor distance analysis reveals that mitotic nucleosome clusters have the same internal packing density as in interphase. Furthermore, mitotic chromosomes contain fewer megacomplexes. This uneven chromosome condensation helps explain a longstanding enigma of mitosis: most genes are repressed but a subset is upregulated.

cell biology↗

bioRxiv2017

Deep Recurrent Neural Network Reveals a Hierarchy of Process Memory during Dynamic Natural Vision

The human visual cortex extracts both spatial and temporal visual features to support perception and guide behavior. Deep convolutional neural networks (CNNs) provide a computational framework to model cortical representation and organization for spatial visual processing, but unable to explain how the brain processes temporal information. To overcome this limitation, we extended a CNN by adding recurrent connections to different layers of the CNN to allow spatial representations to be remembered and accumulated over time. The extended model, or the recurrent neural network (RNN), embodied a hierarchical and distributed model of process memory as an integral part of visual processing. Unlike the CNN, the RNN learned spatiotemporal features from videos to enable action recognition. The RNN better predicted cortical responses to natural movie stimuli than the CNN, at all visual areas especially those along the dorsal stream. As a fully-observable model of visual processing, the RNN also revealed a cortical hierarchy of temporal receptive window, dynamics of process memory, and spatiotemporal representations. These results support the hypothesis of process memory, and demonstrate the potential of using the RNN for in-depth computational understanding of dynamic natural vision.

neuroscience↗

bioRxiv2017

Transferring and Generalizing Deep-Learning-based Neural Encoding Models across Subjects

Recent studies have shown the value of using deep learning models for mapping and characterizing how the brain represents and organizes information for natural vision. However, modeling the relationship between deep learning models and the brain (or encoding models), requires measuring cortical responses to large and diverse sets of natural visual stimuli from single subjects. This requirement limits prior studies to few subjects, making it difficult to generalize findings across subjects or for a population. In this study, we developed new methods to transfer and generalize encoding models across subjects. To train encoding models specific to a subject, the models trained for other subjects were used as the prior models and were refined efficiently using Bayesian inference with a limited amount of data from the specific subject. To train encoding models for a population, the models were progressively trained and updated with incremental data from different subjects. For the proof of principle, we applied these methods to functional magnetic resonance imaging (fMRI) data from three subjects watching tens of hours of naturalistic videos, while deep residual neural network driven by image recognition was used to model the visual cortical processing. Results demonstrate that the methods developed herein provide an efficient and effective strategy to establish subject-specific or population-wide predictive models of cortical representations of high-dimensional and hierarchical visual features.

neuroscience↗

bioRxiv2017

DNA 5-Hydroxymethylcytosines from Cell-free Circulating DNA as Diagnostic Biomarkers for Human Cancers

DNA modifications such as 5-methylcytosines (5mC) and 5-hydroxymethylcytosines (5hmC) are epigenetic marks known to affect global gene expression in mammals(1, 2). Given their prevalence in the human genome, close correlation with gene expression, and high chemical stability, these DNA epigenetic marks could serve as ideal biomarkers for cancer diagnosis. Taking advantage of a highly sensitive and selective chemical labeling technology(3), we report here genome-wide 5hmC profiling in circulating cell-free DNA (cfDNA) and in genomic DNA of paired tumor/adjacent tissues collected from a cohort of 90 healthy individuals and 260 patients recently diagnosed with colorectal, gastric, pancreatic, liver, or thyroid cancer. 5hmC was mainly distributed in transcriptionally active regions coincident with open chromatin and permissive histone modifications. Robust cancer-associated 5hmC signatures in cfDNA were identified with specificity for different cancers. 5hmC-based biomarkers of circulating cfDNA demonstrated highly accurate predictive value for patients with colorectal and gastric cancers versus healthy controls, superior to conventional biomarkers, and comparable to 5hmC biomarkers from tissue biopsies. This new strategy could lead to the development of effective blood-based, minimally-invasive cancer diagnosis and prognosis approaches.

cancer biology↗

bioRxiv2017

Deep Residual Network Reveals a Nested Hierarchy of Distributed Cortical Representation for Visual Categorization

The brain represents visual objects with topographic cortical patterns. To address how distributed visual representations enable object categorization, we established predictive encoding models based on a deep residual neural network, and trained them to predict cortical responses to natural movies. Using this predictive model, we mapped human cortical representations to 64,000 visual objects from 80 categories with high throughput and accuracy. Such representations covered both the ventral and dorsal pathways, reflected multiple levels of object features, and preserved semantic relationships between categories. In the entire visual cortex, object representations were modularly organized into three categories: biological objects, non-biological objects, and background scenes. In a finer scale specific to each module, object representations revealed sub-modules for further categorization. These findings suggest that increasingly more specific category is represented by cortical patterns in progressively finer spatial scales. Such a nested hierarchy may be a fundamental principle for the brain to categorize visual objects with various levels of specificity, and can be explained and differentiated by object features at different levels.

neuroscience↗

bioRxiv2017

Variable sensitivity to DNA damaging chemotherapeutic modulated by cell type-dependent bimodal p53 dynamics

Mechanisms that determine drug sensitivity of distinct cancer types is poorly understood for most cytotoxic chemotherapy. In this study, we elucidated a new resistance mechanism to DNA damaging chemotherapeutic through modulation of p53 dynamics. While both sensitive and resistant cancer cell lines activated similar p53 oscillation followed by cell-cycle arrest in response to low dose of DNA damaging drug, they switched in a bimodal manner to monotonic or single pulse dynamics at high drug dose. Cell lines with monotonically increasing p53 underwent rapid and extensive drug-induced apoptosis, while those exhibiting a single p53 pulse mostly survived. By combining single cell imaging with computational modeling, we characterized a regulatory module involving ATM, p53, Mdm2 and Wip1, which generates bimodal p53 dynamics through coupled feed-forward and feedback, and we found that basal expression of ATM determined the differential modular output between drug sensitive and resistant lines. Moreover, we showed combinatorial inhibition of Mdm2 and Wip1 was an effective strategy to alter p53 dynamics in resistant cancer cells and sensitize their apoptotic response. Our results point to p53 pulsing as a potentially druggable mechanism that mediates resistance to cytotoxic chemotherapy.

systems biology↗

bioRxiv2017

Natural Chromatin Is Heterogeneous And Self Associates In Vitro

The 30-nm fiber is commonly found in oligonucleosome arrays in vitro but rarely found in chromatin within nuclei. To determine how chromatin high-order structure is controlled, we used cryo-ET to study the undigested natural chromatin released from cells that do not have evidence of 30-nm fibers in vivo: picoplankton and yeast. In the presence of divalent cations, most of the chromatin from both organisms is compacted into a large mass. Rare irregular 30-nm fibers do form at the periphery of this mass, some of which include face-to-face interactions. In the absence of divalent cations, picoplankton chromatin decondenses into open zigzags. By contrast, yeast chromatin mostly remains compact with looser nucleosome packing, even after treatment with histone-deacetylase inhibitor. The 3-D configuration of natural chromatin is therefore sensitive to the local environment, but generally nonpermissive of regular motifs, even at the level of oligonucleosomes.

cell biology↗

bioRxiv2017

Uncharacterized bacterial structures revealed by electron cryotomography

SUMMARY STATEMENTHere we present a survey of previously uncharacterized structures we have observed in bacterial cells by electron cryotomography, in the hopes of spurring their identification and study.\n\nABSTRACTElectron cryotomography (ECT) can reveal the native structure and arrangement of macromolecular complexes inside intact cells. This technique has greatly advanced our understanding of the ultrastructure of bacterial cells. Rather than undifferentiated bags of enzymes, we now view bacteria as structurally complex assemblies of macromolecular machines. To date, our group has applied ECT to nearly 90 different bacterial species, collecting more than 15,000 cryotomograms. In addition to known structures, we have observed several, to our knowledge, uncharacterized features in these tomograms. Some are completely novel structures; others expand the features or species range of known structure types. Here we present a survey of these uncharacterized bacterial structures in the hopes of accelerating their identification and study, and furthering our understanding of the structural complexity of bacterial cells.

microbiology↗

bioRxiv2017

Gene expression profiling reveals U1 snRNA regulates cancer gene expression

U1 small nuclear RNA (U1 snRNA), as one of the most abundant noncoding RNA in eukaryotic cells plays an important role in splicing of pre-mRNAs. Compared to other studies which have focused on the primary function of U1 snRNA and the neurodegenerative diseases caused by the abnormalities of U1 snRNA, this study is to investigate how the U1 snRNA over-expression affects the expression of genes on a genome-wide scale. In this study, we built a model of U1 snRNA over-expression in a rat cell line. By comparing the gene expression profiles of U1 snRNA over-expressed cells with those of their controls using the microarray experiments, 916 genes or loci were identified significantly differentially expressed. These 595 up-regulated genes and 321 down-regulated genes were further analyzed using the annotations from the GO terms and the KEGG database. As a result, three of 12 enriched pathways are well-known cancer pathways, while nine of them were associated to cancers in previous studies. The further analysis of 73 genes involved in 12 pathways suggests that U1 snRNA regulates cancer gene expression. The microarray data with ID GSE84304 is available in the NCBI GEO database.

cancer biology↗

bioRxiv2016

The complex sequence landscape of maize revealed by single molecule technologies

Complete and accurate reference genomes and annotations provide fundamental tools for characterization of genetic and functional variation. These resources facilitate elucidation of biological processes and support translation of research findings into improved and sustainable agricultural technologies. Many reference genomes for crop plants have been generated over the past decade, but these genomes are often fragmented and missing complex repeat regions. Here, we report the assembly and annotation of maize, a genetic and agricultural model species, using Single Molecule Real-Time (SMRT) sequencing and high-resolution optical mapping. Relative to the previous reference genome, our assembly features a 52-fold increase in contig length and significant improvements in the assembly of intergenic spaces and centromeres. Characterization of the repetitive portion of the genome revealed over 130,000 intact transposable elements (TEs), allowing us to identify TE lineage expansions unique to maize. Gene annotations were updated using 111,000 full-length transcripts obtained by SMRT sequencing. In addition, comparative optical mapping of two other inbreds revealed a prevalence of deletions in the low gene density region and maize lineage-specific genes.

genomics↗