Wound infection is a serious complication in burn injury, which is a common form of trauma and an important public health issue. We investigated samples from burn and nonburn wounds for microbial characteristics and temporal trends of antibiotic resistance. Wound samples were collected from 369 burned patients and 927 non-burned individuals admitted from 2007 to 2017. Higher frequency of Acinetobacter baumannii, Klebsiella pneumonia, and Pseudomonas aeruginosa was observed in samples from burned individuals when compared to those from non-burned. The prevalence of different groups of bacteria varied when the samples were stratified according to age and sex. The antimicrobial resistance profiles showed a significant difference between burned and non-burned patients. The different temporal trends of antimicrobial resistance rates were also found, which may be critical for the proper selection of antibiotics in burn treatment. The present study suggested that frequent pathogens and antibacterial resistance evolution could differ between burn wounds and other wounds. Therefore, periodic surveillance of antibiotic resistance patterns in the burn unit might help physicians properly select antibiotics for treatment.

Abstract P53 is the ‘guardian of the genome’ and is responsible for regulating cell cycle and apoptosis. The genomic p53 binding regions, where activating transcriptional factors and cofactors like p300 simultaneously bind, are called ‘p53-dependent enhancers’, which play an important role in tumorigenesis. Current experimental assays generally provide a broad peak of each enhancer element, leaving our knowledge about critical enhancer regions (CERs) limited. Under the inspiration of enhancer dissection by CRISPR-Cas9 screen library on genome-wide p53 binding sites, here we introduce a statistical framework called ‘Computational CRISPR Strategy’ (CCS), to predict whether a given DNA fragment will be a p53-dependent CER by employing 7-mer as feature extractions along with random forest as the regressor. When training on a p53 CRISPR enhancer dataset, CCS not only accurately fitted the top-ranked enriched single guide RNAs (sgRNAs) but also successfully reproduced two known CERs that were validated by experiments. When applying it to an independent testing dataset on a tilling of a 2K-b genomic region of CRISPR-deCDKN1A-Lib, the trained model shows great generalizability by identifying a CER containing five top-ranked sgRNAs. A feature importance analysis further indicates that top-ranked 7-mers are mapped onto informative TF motifs including POU5F1 and SOX5, which are differentially enriched in p53-dependent CERs and are potential factors to make a general p53 binding site to form a p53-dependent CER, providing the interpretability of the trained model. Our results demonstrate that CCS is an alternative way of the CRISPR experiment to screen the genome for mapping p53-dependent CERs.

Deciphering the code of cis-regulatory element (CRE) is one of the core issues of today’s biology. Enhancers are distal CREs and play significant roles in gene transcriptional regulation. Although identifications of enhancer locations across the whole genome [discriminative enhancer predictions (DEP)] is necessary, it is more important to predict in which specific cell or tissue types, they will be activated and functional [tissue-specific enhancer predictions (TSEP)]. Although existing deep learning models achieved great successes in DEP, they cannot be directly employed in TSEP because a specific cell or tissue type only has a limited number of available enhancer samples for training. Here, we first adopted a reported deep learning architecture and then developed a novel training strategy named “pretraining-retraining strategy” (PRS) for TSEP by decomposing the whole training process into two successive stages: a pretraining stage is designed to train with the whole enhancer data for performing DEP, and a retraining strategy is then designed to train with tissue-specific enhancer samples based on the trained pretraining model for making TSEP. As a result, PRS is found to be valid for DEP with an AUC of 0.922 and a GM (geometric mean) of 0.696, when testing on a larger-scale FANTOM5 enhancer dataset via a five-fold cross-validation. Interestingly, based on the trained pretraining model, a new finding is that only additional twenty epochs are needed to complete the retraining process on testing 23 specific tissues or cell lines. For TSEP tasks, PRS achieved a mean GM of 0.806 which is significantly higher than 0.528 of gkm-SVM, an existing mainstream method for CRE predictions. Notably, PRS is further proven superior to other two state-of-the-art methods: DEEP and BiRen. In summary, PRS has employed useful ideas from the domain of transfer learning and is a reliable method for TSEPs.

Recognition of DNA-binding protein is a very meaningful work, because DNA-binding proteins act as the very vital roles in many biological processes. In order to reveal the inner connection between intrinsic information of protein and the binding force of DNA and protein, a 314-dimensional vector is inputted for DNA-binding protein prediction. And all the 314 dimensional values are identified as the more vital digital feature and coded from the multiple properties of protein. A large number of mathematical experiments are performed with 5-fold cross validation test to find the optimal parameters and construct the available models with random forest and elastic net. The numeric features of box-counting dimension, information entropies of chaos game representation and information entropies of dipeptide composition are regarded as more crucial roles showed by a large number of experiments. The performance of random forest model and elastic net model of this study is slightly better than the one of DNA-Prot for test dataset. The Matthew's correlation coefficient (MCC) is 0.7374 and 0.7591 and accuracy (ACC) achieves respectively 0.8750 and 0.8698. For independent dataset1 and independent dataset2 it gains slightly lower MCC and ACC value than DNA-Prot [1].

DNA-binding proteins are involved and play a crucial role in a lot of important biological processes. Hence, the identification of the DNA-binding proteins is a challenging and significant problem. In order to reveal the intrinsic information correlated to DNA-binding, nine classes of candidate features based on different mathematical fields are applied to construct the prediction model with random forest. They are fractal dimension, conjoint triad feature, Hilbert-Huang Transformation, amino acid composition, dipeptide composition, chaos game representation, and the corresponding information entropies. These mathematical expressions are evaluated with 5-fold cross validation test. The results of numerical simulations show that the mathematical features consisted of amino acid composition, fractal dimension and information entropies of amino acid and chaos game representation achieve the best performance. Its accuracy is 0.8157, and Matthew's correlation coefficient (MCC) achieves 0.5968 on the benchmark dataset from DNA-Prot. By analyzing the components of top combination of the nine candidate features, the concepts of fractal dimension and information entropy are the effective and vital features, which can provide complementary sequence-order information on the basis of amino acid composition.