Abstract According to the World Health Organization, thousands of people die every year in road traffic accidents. A crucial problem is the prediction of medical assistance in these accidents. For this purpose, we propose a new deep learning model whose goal is to distinguish whether a traffic accident requires medical assistance. The proposed perspective is general, so the model is valid for any dataset from any city. For this purpose, we present a model divided into three differentiated stages. In the first pre-processing stage, a general data treatment is performed, from data collection and cleaning to balancing. Secondly, the post-processing stage employs genetic and boosting algorithms to obtain the importance of all the data set variables used in the prediction. In the last stage, Model Training, a new model based on two-dimensional convolutional neural networks is applied to obtain a prediction of the need for medical assistance in traffic accidents. Finally, we test the effectiveness and accuracy of the proposed model by applying it to traffic accident datasets in six different cities. The obtained experimental results show that our framework achieves higher accuracy in all cities compared to six state-of-the-art models, confirming its suitability and applicability, even in real time.

BACKGROUND: Despite their involvement in the regulation of gene expression and their importance as genomic markers for promoter prediction, no objective standard exists for defining CpG islands (CGIs), since all current approaches rely on a large parameter space formed by the thresholds of length, CpG fraction and G+C content. RESULTS: Given the higher frequency of CpG dinucleotides at CGIs, as compared to bulk DNA, the distance distributions between neighboring CpGs should differ for bulk and island CpGs. A new algorithm (CpGcluster) is presented, based on the physical distance between neighboring CpGs on the chromosome and able to predict directly clusters of CpGs, while not depending on the subjective criteria mentioned above. By assigning a p-value to each of these clusters, the most statistically significant ones can be predicted as CGIs. CpGcluster was benchmarked against five other CGI finders by using a test sequence set assembled from an experimental CGI library. CpGcluster reached the highest overall accuracy values, while showing the lowest rate of false-positive predictions. Since a minimum-length threshold is not required, CpGcluster can find short but fully functional CGIs usually missed by other algorithms. The CGIs predicted by CpGcluster present the lowest degree of overlap with Alu retrotransposons and, simultaneously, the highest overlap with vertebrate Phylogenetic Conserved Elements (PhastCons). CpGcluster's CGIs overlapping with the Transcription Start Site (TSS) show the highest statistical significance, as compared to the islands in other genome locations, thus qualifying CpGcluster as a valuable tool in discriminating functional CGIs from the remaining islands in the bulk genome. CONCLUSION: CpGcluster uses only integer arithmetic, thus being a fast and computationally efficient algorithm able to predict statistically significant clusters of CpG dinucleotides. Another outstanding feature is that all predicted CGIs start and end with a CpG dinucleotide, which should be appropriate for a genomic feature whose functionality is based precisely on CpG dinucleotides. The only search parameter in CpGcluster is the distance between two consecutive CpGs, in contrast to previous algorithms. Therefore, none of the main statistical properties of CpG islands (neither G+C content, CpG fraction nor length threshold) are needed as search parameters, which may lead to the high specificity and low overlap with spurious Alu elements observed for CpGcluster predictions.

MOTIVATION: DNA sequences are formed by patches or domains of different nucleotide composition. In a few simple sequences, domains can simply be identified by eye; however, most DNA sequences show a complex compositional heterogeneity (fractal structure), which cannot be properly detected by current methods. Recently, a computationally efficient segmentation method to analyse such nonstationary sequence structures, based on the Jensen-Shannon entropic divergence, has been described. Specific algorithms implementing this method are now needed. RESULTS: Here we describe a heuristic segmentation algorithm for DNA sequences, which was implemented on a Windows program (SEGMENT). The program divides a DNA sequence into compositionally homogeneous domains by iterating a local optimization procedure at a given statistical significance. Once a sequence is partitioned into domains, a global measure of sequence compositional complexity (SCC), accounting for both the sizes and compositional biases of all the domains in the sequence, is derived. SEGMENT computes SCC as a function of the significance level, which provides a multiscale view of sequence complexity.

An algorithm to simulate DNA sequence evolution under a general stochastic model, including as particular cases all the previously used schemes of nucleotide substitution, is described. The stimulation is carried out on finite, variable length, DNA sequences through a strict stochastic process, according to the particular substitution rates imposed by each scheme. Five FORTRAN programs, running on an IBM PC and compatibles, carry out all the tasks needed for the simulation. They are menu driven and interfaced to the system through a principal menu. All sequence data files used and generated by the SDSE package conform to the standard GenBank database format, thus allowing the use of any sequence retrieved from this databank, as well as the application of other packages to analyse, manipulate or retrieve stimulated sequences.