Hi-C is commonly used to study three-dimensional genome organization. However, due to the high sequencing cost and technical constraints, the resolution of most Hi-C datasets is coarse, resulting in a loss of information and biological interpretability. Here we develop DeepHiC, a generative adversarial network, to predict high-resolution Hi-C contact maps from low-coverage sequencing data. We demonstrated that DeepHiC is capable of reproducing high-resolution Hi-C data from as few as 1% downsampled reads. Empowered by adversarial training, our method can restore fine-grained details similar to those in high-resolution Hi-C matrices, boosting accuracy in chromatin loops identification and TADs detection, and outperforms the state-of-the-art methods in accuracy of prediction. Finally, application of DeepHiC to Hi-C data on mouse embryonic development can facilitate chromatin loop detection. We develop a web-based tool (DeepHiC, http://sysomics.com/deephic) that allows researchers to enhance their own Hi-C data with just a few clicks. We developed a novel method, DeepHiC, for enhancing Hi-C data resolution from low-coverage sequencing data using generative adversarial network. DeepHiC is capable of reproducing high-resolution (10-kb) Hi-C data with high quality even using 1/100 downsampled reads. Our method outperforms the previous methods in Hi-C data resolution enhancement, boosting accuracy in chromatin loops identification and TADs detection. Application of DeepHiC on mouse embryonic development data shows that enhancements afforded by DeepHiC facilitates the chromatin loops prediction of these data, yielding significant interactions more enriched in open chromatin regions and gene promoters. We also developed a user-friendly web server (http://sysomics.com/deephic) that allows researchers to enhance their own low-coverage Hi-C data with just few clicks.

Entity linking refers to the task of aligning mentions of entities in the text to their corresponding entries in a specific knowledge base, which is of great significance for many natural language process applications such as semantic text understanding and knowledge fusion. The pivotal of this problem is how to make effective use of contextual information to disambiguate mentions. Moreover, it has been observed that, in most cases, mention has similar or even identical strings to the entity it refers to. To prevent the model from linking mentions to entities with similar strings rather than the semantically similar ones, in this paper, we introduce the advanced language representation model called BERT (Bidirectional Encoder Representations from Transformers) and design a hard negative samples mining strategy to fine-tune it accordingly. Based on the learned features, we obtain the valid entity through computing the similarity between the textual clues of mentions and the entity candidates in the knowledge base. The proposed hard negative samples mining strategy benefits entity linking from the larger, more expressive pre-trained representations of BERT with limited training time and computing sources. To the best of our knowledge, we are the first to equip entity linking task with the powerful pre-trained general language model by deliberately tackling its potential shortcoming of learning literally, and the experiments on the standard benchmark datasets show that the proposed model yields state-of-the-art results.

Modern cybersecurity built on public-key cryptosystems like Rivest-Shamir-Adleman is compromised upon finding solutions to the prime factorization. Nevertheless, solving the prime factorization problem, given a large N , remains computationally challenging. Here, we design DNA origami frameworks (DOFs) to direct localized assembly of double-crossover (DX) tiles for solving prime factorization with a model consisting of the computing, decision-making, and reporting motifs. The model implementation is based on the sequential assembly of different DX tiles in the DOF cavity that carries overhangs encoding the prime and composite integers. The primes are multiplied and then verified with the composite, and the result is visualized under atomic force microscopy via the presence (success) or absence (failure) of biotin-streptavidin labels on the reporting DX tile. The factorization of semiprimes 6 and 15 is realized with this DOF-based demonstration. Given the potential of massively parallel processing ability of DNA, this strategy opens an avenue to solve complex mathematical puzzles like prime factoring with molecular computing.

With accumulating public omics data, great efforts have been made to characterize the genetic heterogeneity of breast cancer. However, identifying novel targets and selecting the best from the sizeable lists of candidate targets is still a key challenge for targeted therapy, largely owing to the lack of economical, efficient and systematic discovery and assessment to prioritize potential therapeutic targets. Here, we describe an approach that combines the computational evaluation and objective, multifaceted assessment to systematically identify and prioritize targets for biological validation and therapeutic exploration. We first establish the reference gene expression profiles from breast cancer cell line MCF7 upon genome-wide RNA interference (RNAi) of a total of 3689 genes, and the breast cancer query signatures using RNA-seq data generated from tissue samples of clinical breast cancer patients in the Cancer Genome Atlas (TCGA). Based on gene set enrichment analysis, we identified a set of 510 genes that when knocked down could significantly reverse the transcriptome of breast cancer state. We then perform multifaceted assessment to analyze the gene set to prioritize potential targets for gene therapy. We also propose drug repurposing opportunities and identify potentially druggable proteins that have been poorly explored with regard to the discovery of small-molecule modulators. Finally, we obtained a small list of candidate therapeutic targets for four major breast cancer subtypes, i.e., luminal A, luminal B, HER2+ and triple negative breast cancer. This RNAi transcriptome-based approach can be a helpful paradigm for relevant researches to identify and prioritize candidate targets for experimental validation.

Entity attribute extraction, which converts chaotic text data to structured knowledge, plays an important role in natural language processing (NLP). Many previous studies proved that the representation of text has a significant impact on the results of attribute extraction. In this paper, we propose a novel model to obtain the discriminative representation of sentences by applying the Siamese architecture. Specifically, we simultaneously input two variable-length sentences in the training stage of our model, namely, the main sentence and its similar or dissimilar partner. This scheme of pair is beneficial for entity attribute extraction. First, the entity attribute extraction community suffers the insufficient but expensive labeled data, the two input sets produce much more samples for the representation learning and can be treated as a useful data augmentation method. More importantly, the co-learning by the Siamese architecture achieves more interesting embedding than the separate way, since the informative relation between the focused sentence and its partner help the representation learning to explore more essential semantics and keep stable to the variation of wording and syntax of a sentence. The experiments on the Wikipedia data show that our model takes advantage of the Siamese architecture for sentence embedding and achieves significant improvements on attribute extraction as compared with baselines. To the best of our knowledge, we are the first to introduce the Siamese network into the person entity attribute extraction, which we proved to achieve the state of the art.

Radioresistance represents the predominant cause for radiotherapy failure and disease progression, resulting in increased breast cancer mortality. Through gene expression signatures analyses of Library of Integrated Network-Based Cellular Signatures (LINCS) and Gene Expression Omnibus (GEO), the present study aimed to identify potential candidate radiosensitizers from known drugs systematically. The similarity of integrated gene expression signatures between irradiated eIF4G1 -silenced breast cancer cells and known drugs was measured by enrichment scores. Drugs with positive enrichment scores were selected as potential radiosensitizers. The radiosensitizing effects of the candidate radiosensitizers were analyzed in breast cancer cells (MCF-7, MX-1, and MDA-MB-231) by CCK-8 assays and colony-forming capability after exposure to ionizing radiation. Cell apoptosis was detected by flow cytometry. Expressions of eIF4G1 and a series of DNA damage response proteins were analyzed by Western blot assays. Bosutinib was proposed to be a promising radiosensitizer as its administration markedly reduced the dosages of both the drug and ionizing radiation and was associated with fewer adverse drug reactions. The combined treatment with ionizing radiation and bosutinib significantly increased the cells killing potency in all three cell lines as compared to ionizing radiation or bosutinib alone. MX-1 cells were revealed to be the most sensitive to both ionizing radiation and bosutinib among the three cell lines. Bosutinib noticeably downregulated the expression of eIF4G1 in a dose-dependent manner and also reduced the expression of DNA damage response proteins (including ATM , XRCC4 , ATRIP , and GADD45a ). Moreover, eIF4G1 could be a key target of bosutinib through which it regulates DNA damage induced by ionizing radiation. Thus, taken together, bosutinib may serve as a potential candidate radiosensitizer for breast cancer therapy.