Each Department of Transportation in the United States must provide to the Federal Highway Administration on annual basis the number and types of vehicles traveled on its state-maintained roads. These data are fed into the Highway Performance Monitoring System used to assess the nation’s highway system performance. Classifying vehicles (i.e., identifying their types, e.g., passenger cars, trucks, etc.) during nighttime is quite challenging due to limited lighting. This study designed and evaluated three Convolutional Neural Network (CNN) models to classify vehicles using their thermal images. These three models have architectures that differ in the number of layers and, in the case of the third model, the addition of an inception layer. Of these, the second model achieves the best performance, achieving mean accuracy scores of greater than 97% for each of the three vehicle classes and f1 scores of greater than 98%. We proposed two training-test methods based on data augmentation to avoid over-fitting and to improve performance. The experimental results demonstrated that a data augmentation training-test method improves model performance further with regard to both accuracy and f1-score.

Proteotypic peptides are the peptides in protein sequences that can be confidently observed by mass-spectrometry based proteomics. In recent years, there has been an increased effort to use proteotypic peptide prediction to improve the accuracy of peptide identification. These investigations compile various physicochemical peptide features to identify whether peptides are proteotypic. Here we describe our method for the selection, reduction and evaluation of physicochemical features for proteotypic peptide prediction. We performed feature selection on a published set of features and identified six features as the most significant. To highlight the effectiveness of our reduced feature set, we trained three machine learning algorithms (support vector machines, random forests, and XGBoost) as proteotypic peptide identifiers. Importantly, for larger data sets, the random forests and XGBoost algorithms trained faster than the support vector machine, as solving the support vector machine objective function requires quadratic programming. Our three classifiers had similar if not better prediction accuracy when compared to other proteotypic peptide predictors on the same data sets.

In this work, we investigate privacy violations that occur when non-confidential medical data is combined with domain ontologies to infer confidential data. We propose enhancements to our existing framework to detect such privacy violations, and to eliminate undesired inferences in an efficient manner. Our enhanced inference channel removal methods are based on disruption covers and heuristic-guided modification of data items that contribute to an inference. We show that our method is sound and complete. Soundness means that we modify only data items that lead to undesired inferences. Completeness means that we remove all inferences leading to undesired data disclosures. Finally, we show that our solution is practical with respect to computational cost. An important aspect of our approach is that it sets the foundation for creating patient-specific privacy policies; an emerging need in the healthcare domain.

In this paper, we present HPMA, a graphics processing unit (GPU) accelerated meta-genome sequence alignment algorithm for a collection of DNA sequences. This algorithm supports all-to-all pairwise local alignment on NVIDIA GPUs. HPMA builds on an GPU alignment algorithm that we developed earlier with the addition of a filter module. We designed and developed this new kernel function based on the suffix array data structure. The filter module improves performance by identifying a subset of sequences which meet a user-defined similarity threshold and should be considered for alignment. HPMA has the ability to balance the workload between CPU and GPU. HPMA allows us to preprocess massively large metagenomes in a reasonable amount of time in response to increasing speed of NGS sequencers. The performance of HPMA has been evaluated on a cluster of Kepler-based Tesla K20 GPUs using a variety of short DNA sequence datasets. We evaluate HPMA thoroughly with four test datasets. The first two test sets are comprised of 10 simulated datasets where read length varies from 72 to 750 base-pairs. The third test set is designed to allow a comparison with published results for GSWABE, a competing GPU alignment tool. The fourth test set is an actual metagenome of over 2 million sequences with an average length of 270 bp. We utilized a cluster of NVIDIA-K20 GPUs in the Stampede supercomputer at the Texas Advanced Computing Center (Austin, TX, USA). When running on a cluster of 10 NVIDIA K20 GPUs, HPMA is able to align 2 million simulated metagenome sequences of length 300 bp in 160 seconds. In the case of real metagenomic data, HPMA is able to align 2,038,516 sequences with an average length of 270 bp in 60 seconds.

BACKGROUND: Computational gene finding algorithms have proven their robustness in identifying genes in complete genomes. However, metagenomic sequencing has presented new challenges due to the incomplete and fragmented nature of the data. During the last few years, attempts have been made to extract complete and incomplete open reading frames (ORFs) directly from short reads and identify the coding ORFs, bypassing other challenging tasks such as the assembly of the metagenome. RESULTS: In this paper we introduce a metagenomics gene caller (MGC) which is an improvement over the state-of-the-art prediction algorithm Orphelia. Orphelia uses a two-stage machine learning approach and computes a model that classifies extracted ORFs from fragmented sequences. We hypothesise and demonstrate evidence that sequences need separate models based on their local GC-content in order to avoid the noise introduced to a single model computed with sequences from the entire GC spectrum. We have also added two amino-acid features based on the benefit of amino-acid usage shown in our previous research. Our algorithm is able to predict genes and translation initiation sites (TIS) more accurately than Orphelia which uses a single model. CONCLUSIONS: Learning separate models for several pre-defined GC-content regions as opposed to a single model approach improves the performance of the neural network as demonstrated by the experimental results presented in this paper. The inclusion of amino-acid usage features also helps improve the overall accuracy of our algorithm. MGC's improvement sets the ground for further investigation into the use of GC-content to separate data for training models in machine learning based gene finders.

The effectiveness of de novo peptide sequencing algorithms depends on the quality of MS/MS spectra. Since most of the peaks in a spectrum are uninterpretable `noise' peaks it is necessary to carefully pre-filter the spectra to identify the `signal' peaks that likely correspond to b-/y-ions. Selecting the optimal set of peaks for candidate peptide generation is essential for obtaining accurate results. A careful balance must be maintained between the precision and recall of peaks that are selected for further processing and candidate peptide generation. If too many peaks are selected the search space will be too large and the problem becomes intractable. If too few peaks are selected cleavage sites will be missed, the resulting candidate peptides will have large gaps, and sequencing results will be poor. For this reason pre-filtering of MS/MS spectra and accurate selection of peaks for peptide candidate generation is essential to any de novo peptide sequencing algorithm. We present a novel neural network approach for the selection of b-/y-ions using known fragmentation characteristics, and leveraging neural network probability estimates of flanking and complementary ions. We show a significant improvement in precision and recall of peaks corresponding to b-/y-ions and a reduction in search space over approaches used by other de novo peptide sequencing algorithms.

MOTIVATION: In the event of an outbreak of a disease caused by an initially unknown pathogen, the ability to characterize anonymous sequences prior to isolation and culturing of the pathogen will be helpful. We show that it is possible to classify viral sequences by genome type (dsDNA, ssDNA, ssRNA positive strand, ssRNA negative strand, retroid) using amino acid distribution. RESULTS: In this paper we describe the results of analysis of amino acid preference in mammalian viruses. The study was carried out at the genome level as well as two shorter sequence levels: short (300 amino acids) and medium length (660 amino acids). The analysis indicates a correlation between the viral genome types dsDNA, ssDNA, ssRNA positive strand, ssRNA negative strand and retroid and amino acid preference. We investigated three different models of amino acid preference. The simplest amino acid preference model, 1-AAP, is a normalized description of the frequency of amino acids in genomes of a viral genome type. A slightly more complex model is the ordered pair amino acid preference model (2-AAP), which characterizes genomes of different viral genome types by the frequency of ordered pairs of amino acids. The most complex and accurate model is the ordered triple amino acid preference model (3-AAP), which is based on ordered triples of amino acids. The results demonstrate that mammalian viral genome types differ in their amino acid preference. AVAILABILITY: The tools used to format and analyze data and supplementary material are available at http://www.cse.sc.edu/~rose/aminoPreference/index.html CONTACT: [email protected].

UNLABELLED: Dblox and RDblox provide a simple statistical test for duplicated genomic structure; the same programs can also be used to identify putatively duplicated regions. The method focuses on ancient duplication events involving protein-coding genes. AVAILABILITY: http://www.biol.sc.edu/~austin/

Article Free Access Share on Knowledge discovery in reaction databases Authors: J. Royce Rose Department of Computer Science, State University of N. Y. at Stony Brook, Stony Brook, New York Department of Computer Science, State University of N. Y. at Stony Brook, Stony Brook, New YorkView Profile , Herbert Gelernter Department of Computer Science, State University of N. Y. at Stony Brook, Stony Brook, New York Department of Computer Science, State University of N. Y. at Stony Brook, Stony Brook, New YorkView Profile Authors Info & Claims CIKM '93: Proceedings of the second international conference on Information and knowledge managementDecember 1993 Pages 714–716https://doi.org/10.1145/170088.170464Published:01 December 1993Publication History 0citation184DownloadsMetricsTotal Citations0Total Downloads184Last 12 Months2Last 6 weeks2 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteeReaderPDF

Solo is a portable window interface written in the Common Lisp Object System (CLOS) object-oriented programming language. Solo provides a virtual window machine which is targeted to a host window system by implementing a set of host window system specific classes and methods for Solo's host window system driver protocol. The interface presented by Solo to an application insulates it from differences in the host window system, facilitating application portability. Solo distinguishes itself from other object-oriented window systems by exploiting certain features of CLOS. CLOS method combination simplifies initialization of windows while preserving easy extensibility of the basic classes. Generic dispatch on multiple arguments, a feature unique to CLOS, allows a simpler and more flexible input event dispatching protocol. A powerful event description language simplifies the specification of keyboard and mouse events. A prototype implementation runs on the server based XII and NeWS host systems, and on the frame buffer based Lucid Window Toolkit.

Article Fast dispatch mechanisms for stock hardware Share on Author: John R. Rose Sun Microsystems, Inc., 2550 Garcia Avenue, Mountain View, CA Sun Microsystems, Inc., 2550 Garcia Avenue, Mountain View, CAView Profile Authors Info & Claims OOPSLA '88: Conference proceedings on Object-oriented programming systems, languages and applicationsJanuary 1988 Pages 27–35https://doi.org/10.1145/62083.62087Online:01 January 1988Publication History 31citation346DownloadsMetricsTotal Citations31Total Downloads346Last 12 Months7Last 6 weeks0 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access

Article Free Access Share on Refined types: highly differentiated type systems and their use in the design of intermediate languages Author: J. R. Rose Thinking Machines Corporation, Cambridge, MA Thinking Machines Corporation, Cambridge, MAView Profile Authors Info & Claims PLDI '88: Proceedings of the ACM SIGPLAN 1988 conference on Programming language design and implementationJune 1988Pages 278–287https://doi.org/10.1145/53990.54018Published:01 June 1988Publication History 0citation312DownloadsMetricsTotal Citations0Total Downloads312Last 12 Months23Last 6 weeks2 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteeReaderPDF