Xiang Zhang

Bio: Xiang Zhang is an academic researcher from Baylor College of Medicine. The author has contributed to research in topics: Medicine & Computer science. The author has an hindex of 154, co-authored 1733 publications receiving 117576 citations. Previous affiliations of Xiang Zhang include University of California, Berkeley & University of Texas MD Anderson Cancer Center.

Giant suppression of photobleaching for single molecule detection via the Purcell effect.

Abstract: We report giant suppression of photobleaching and a prolonged lifespan of single fluorescent molecules via the Purcell effect in plasmonic nanostructures. The plasmonic structures enhance the spontaneous emission of excited fluorescent molecules, reduce the probability of activating photochemical reactions that destroy the molecules, and hence suppress the bleaching. Experimentally, we observe up to a 1000-fold increase in the total number of photons that we can harvest from a single fluorescent molecule before it bleaches. This approach demonstrates the potential of using the Purcell effect to manipulate photochemical reactions at the subwavelength scale.

An engineered exosome for delivering sgRNA:Cas9 ribonucleoprotein complex and genome editing in recipient cells.

Abstract: CRISPR-Cas9 is a versatile genome-editing technology that is a promising gene therapy tactic. However, the delivery of CRISPR-Cas9 is still a major obstacle to its broader clinical application. Here, we confirm that the components of CRISPR-Cas9—sgRNA and Cas9 protein—can be packaged into exosomes, where sgRNA and Cas9 protein exist as a sgRNA:Cas9 ribonucleoprotein complex. Although exosomal CRISPR-Cas9 components can be delivered into recipient cells, they are not adequate to abrogate the target gene in recipient cells. To solve this, we engineered a functionalized exosome (M-CRISPR-Cas9 exosome) that could encapsulate CRISPR-Cas9 components more efficiently. To improve the loading efficiency of Cas9 proteins into exosomes, we artificially engineered exosomes by fusing GFP and GFP nanobody with exosomal membrane protein CD63 and Cas9 protein, respectively. Therefore, Cas9 proteins could be captured selectively and efficiently loaded into exosomes due to the affinity of GFP-GFP nanobody rather than random loading. sgRNA and Cas9 protein exist as a complex in functionalized exosomes and can be delivered into recipient cells. To show the function of modified exosomes-delivered CRISPR-Cas9 components in recipient cells visually, we generated a reporter cell line (A549stop-DsRed) that produced a red fluorescent signal when the stop element was deleted by the sgRNA-guided endonuclease. Using A549stop-DsRed reporter cells, we showed that modified exosomes loaded with CRISPR-Cas9 components abrogated the target gene more efficiently in recipient cells. Our study reports an alternative tactic for CRISPR-Cas9 delivery.

Building pattern recognition in topographic data: examples on collinear and curvilinear alignments

Abstract: Building patterns are important features that should be preserved in the map generalization process. However, the patterns are not explicitly accessible to automated systems. This paper proposes a framework and several algorithms that automatically recognize building patterns from topographic data, with a focus on collinear and curvilinear alignments. For both patterns two algorithms are developed, which are able to recognize alignment-of-center and alignment-of-side patterns. The presented approach integrates aspects of computational geometry, graph-theoretic concepts and theories of visual perception. Although the individual algorithms for collinear and curvilinear patterns show great potential for each type of the patterns, the recognized patterns are neither complete nor of enough good quality. We therefore advocate the use of a multi-algorithm paradigm, where a mechanism is proposed to combine results from different algorithms to improve the recognition quality. The potential of our method is demonstrated by an application of the framework to several real topographic datasets. The quality of the recognition results are validated in an expert survey.

Abstract GS5-05: Resistance to neoadjuvant chemotherapy in triple negative breast cancer mediated by a reversible drug-tolerant state

Abstract: Approximately 50% of patients with localized triple negative breast cancer (TNBC) have substantial residual cancer burden following treatment with neoadjuvant chemotherapy (NACT), resulting in distant metastasis and death for most of these patients. While genomic and phenotypic intra-tumor heterogeneity are pervasive features of TNBCs at the time of diagnosis, the functional contributions of heterogeneous tumor cell populations to chemoresistance have not been elucidated. To investigate tumor evolution accompanying NACT, we employed orthotopic patient-derived xenograft (PDX) models of treatment-naive TNBC, which retain intra-tumor heterogeneity characteristic of human TNBC. We discovered that some PDX models initially exhibited partial sensitivity to standard front-line NACT (Adriamycin plus Cytoxan, AC). Following AC, residual tumors were resistant to chemotherapy but repopulated tumors with chemo-sensitive cells if left untreated, indicating that tumor cells possessed inherent plasticity. To identify the tumor cell subpopulation(s) conferring chemoresistance, we conducted barcode-mediated clonal tracking in three independent PDX models by introducing a high-complexity pooled lentiviral barcode library into PDX tumor cells which were then orthotopically engrafted into recipient mice. Strikingly, residual tumors maintained the same heterogeneous clonal architecture as naive tumors. Concordantly, whole-exome sequencing revealed conservation of genomic subclonal architecture throughout treatment. These results were corroborated by genomic sequencing of serial biopsies pre- and post-AC obtained directly from TNBC patients enrolled on an ongoing clinical trial at MD Anderson (ARTEMIS; NCT02276443). Together, these studies revealed that genomically distinct pre-treatment subclones were equally capable of surviving AC to reconstitute tumors after treatment. To identify functional addictions of residual tumor cells, we conducted histologic and transcriptomic profiling. Residual tumors following AC-treatment exhibited extensive fibrotic desmoplasia and tumor cell pleomorphism in both PDX models and in serial biopsies obtained from TNBC patients enrolled on the ARTEMIS trial. Strikingly, these AC-induced features were reverted upon regrowth of residual tumors in PDXs and in patients9 tumors. Similarly, residual tumors exhibited unique transcriptomic features, many of which are also de-regulated in cohorts of human TNBCs undergoing chemotherapy treatment. These features were nearly completely reverted after tumors regrew, suggesting that the residual tumor state may be a unique and transient therapeutic window. Gene set enrichment analyses revealed that residual tumors had increased activation of oxidative phosphorylation and decreased glycolytic signaling. Pharmacologic targeting of oxidative phosphorylation with a small-molecule inhibitor of mitochondrial electron transport chain complex I (IACS-010759) significantly delayed the regrowth of AC-treated residual tumors in three independent PDX models. Collectively, these studies reveal that a reversible phenotypic state can confer chemoresistance in the absence of genomic selection and that the residual tumor state is a novel therapeutic window for chemo-refractory TNBC. Citation Format: Echeverria GV, Ge Z, Seth S, Jeter-Jones SL, Zhang X, Zhou X, Cai S, Tu Y, McCoy A, Peoples M, Lau R, Shao J, Sun Y, Bristow C, Carugo A, Ma X, Harris A, Wu Y, Moulder S, Symmans WF, Marszalek JR, Heffernan TP, Chang JT, Piwnica-Worms H. Resistance to neoadjuvant chemotherapy in triple negative breast cancer mediated by a reversible drug-tolerant state [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr GS5-05.

FREEDOM-7: A High Fidelity Seven Axis Haptic Device with Application to Surgical Training

Abstract: A seven axis haptic device, called the Freedom-7, is described in relation to its application to surgical training. The generality of its concept makes it also relevant to most other haptic applications. The design rationale is driven by a long list of requirements since such a device is meant to interact with the human hand: uniform response, balanced inertial properties, static balancing, low inertia, high frequency response, high resolution, low friction, arbitrary reorientation, and low visual intrusion. Some basic performance figures are also reported.

Deep Residual Learning for Image Recognition

Abstract: Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers—8× deeper than VGG nets [40] but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions1, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

Very Deep Convolutional Networks for Large-Scale Image Recognition

Abstract: In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks of increasing depth using an architecture with very small (3x3) convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers. These findings were the basis of our ImageNet Challenge 2014 submission, where our team secured the first and the second places in the localisation and classification tracks respectively. We also show that our representations generalise well to other datasets, where they achieve state-of-the-art results. We have made our two best-performing ConvNet models publicly available to facilitate further research on the use of deep visual representations in computer vision.

Deep Residual Learning for Image Recognition

Abstract: Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers---8x deeper than VGG nets but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

Going deeper with convolutions

Abstract: We propose a deep convolutional neural network architecture codenamed Inception that achieves the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. By a carefully crafted design, we increased the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC14 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.