Selected Articles.docx

Use the following examples as a guide (note: abbreviate "Figure" as "Fig." when in the middle of a sentence): "Table 1 provides a selected subset of the most active compounds. The entire list of 96 compounds can be found as Supplementary Table S1 online." "The biosynthetic pathway of L-ascorbic acid in animals involves intermediates of the D-glucuronic acid pathway (see Supplementary Fig. S2 online). Figure 2 shows...".

Selected articles.docx

Please supply all photographic and bitmap images in a bitmap image format such as tiff, jpg, or psd. If saving tiff files, please ensure that the compression option is selected to avoid very large file sizes. Please do not supply Word or Powerpoint files with placed images. Images can be supplied as RGB or CMYK (note: we will not convert image colour modes).

Submit comments on previously published papers, replies to such comments, and all errata to the ECS journal in which the paper commented upon or corrected was published. Comments, replies, and errata article types can be selected in the first stage of the submission process.

Beginning in May 2013, the number of records retrieved from each search for each database was recorded at the moment of searching. The complete results from all databases used for each of the systematic reviews were imported into a unique EndNote library upon search completion and saved without deduplication for this research. The researchers that requested the search received a deduplicated EndNote file from which they selected the references relevant for inclusion in their systematic review. All searches in this study were developed and executed by W.M.B.

We searched PubMed in July 2016 for all reviews published since 2014 where first authors were affiliated to Erasmus MC, Rotterdam, the Netherlands, and matched those with search registrations performed by the medical library of Erasmus MC. This search was used in earlier research [21]. Published reviews were included if the search strategies and results had been documented at the time of the last update and if, at minimum, the databases Embase, MEDLINE, Cochrane CENTRAL, Web of Science, and Google Scholar had been used in the review. From the published journal article, we extracted the list of final included references. We documented the department of the first author. To categorize the types of patient/population and intervention, we identified broad MeSH terms relating to the most important disease and intervention discussed in the article. We copied from the MeSH tree the top MeSH term directly below the disease category or, in to case of the intervention, directly below the therapeutics MeSH term. We selected the domain from a pre-defined set of broad domains, including therapy, etiology, epidemiology, diagnosis, management, and prognosis. Lastly, we checked whether the reviews described limiting their included references to a particular study design.

When comparing to an existing prediction tool, the selected models trained with hybrid features provided a promising accuracy on an independent testing dataset. In short, this work not only characterized the carbonylated substrate preference, but also demonstrated that the proposed method could provide a feasible means for accelerating preliminary discovery of protein carbonylation.

With an attempt to identify useful features for the prediction of protein carbonylation sites, the predictive power of each feature is evaluated based on cross-validation. Additionally, a hybrid approach is investigated in this work by combining different sets of feature vectors with the goal of improving predictive performance on the calssification between carbonylated and non-carbonylated sites. Prior to classification, the data needed to be scaled in the range of [-1, 1] to enhance the effectiveness of prediction [53]. For the construction of predictive models, hybrid features were generated by combining two or more single features. In order to obtain the highest predictive accuracy, the single features were selected based on the mRMR (minimum-redundancy maximum-relevance) [54] algorithm, which sorts the features according to their relevance to the target and the redundancy among the investigated features. The training feature with a smaller index implicates that it has a better trade-off between the maximum relevance and minimum redundancy [16]. The scoring function is defined as follows:

In the investigation of predictive power of single features, the models trained with PWM usually provided better sensitivity than that trained with other features. On the other hand, the models trained with the selected physicochemical properties, top ten AAindices ranked by F-score measurement, could provide best specificity in discriminating carbonylation and non-carbonylation sites. In order to obtain better predictive power, moreover, the models trained with the combination of hybrid features were also evaluated by five-fold cross-validation. The combination of hybrid features was generated by combining two or more single features based on the mRMR-SFS feature-selection method, which incorporates the features sorted by mRMR scores. As presented in Additional file 5: Figure S3, a two-layered predictive model was generated from hybrid features based on mRMR-SFS feature selection. Using SVM as the classifier in Additional file 5: Figure S3), each selected feature was inputted to first-layered SVM for obtaining a feature-specific probability to form an input vector for generating second-layered SVM. In this investigation, the process of feature selection was terminated until predictive performance is not improved anymore. Finally, the models trained with the hybrid features and containing the best cross-validation performance were further evaluated using independent testing datasets.

In classifying between carbonylation and non-carbonylation sites, there is a possibility to overestimate the constructed model due to an overfitting of the training dataset. Thus to evaluate the real performance of the selected models with best cross-validation results, an independent testing dataset was manually extracted from seven research articles, which comprised experimentally verified carbonylation sites from multiple species. As given in Table 7, in classification between 78 carbonylated and 301 non-carbonylated K residues, the SVM model generated using the combination of PWM, AAC and AAindex features provides 0.641, 0.664, 0.659 and 0.252 for sensitivity, specificity, accuracy and MCC value, respectively. The SVM model trained with the hybrid features (PWM, AAindex and AAPC) could give a higher specificity (0.725) in discriminating between 67 carbonylated and 276 non-carbonylated R residues, with the sensitivity of 0.672, the accuracy of 0.714 and MCC value of 0.329. However, the SVM model trained using PWM and AAindex features provides a significantly higher sensitivity (0.755) in carbonylated T residues of the independent testing dataset, while the specificity is slightly low with the value of 0.605. The RF model trained with the the hybrid features (PWM, AAC and AAindex) also achieves a remarkably higher sensitivity (0.755) in carbonylated P residues of the independent testing dataset. In comparison with an existing prediction tool, the CarSPred could provide the best sensitivity (0.811) in carbonylated T residues of the independent testing dataset. Overall, our method performs better than CarSPred based on the independent testing performance.

Do your students want to take part in a Martian mission? Use our Mars scavenger hunt in your next lesson! Students can complete the scavenger hunt activity by reading the selected articles on the NASA Space Place website to find the answers to each clue. Once they have all the clues, they will be able to spell the secret word! You can check their answers with the answer key provided below. 041b061a72