Search

There is a growing desire to create computer systems that can collaborate with humans on complex, open-ended activities. These activities typically have no set completion criteria and frequently involve multimodal communication, extensive world knowledge, creativity, and building structures or compositions through multiple steps. Because these systems differ from question and answer (Q&A) systems, chatbots, and simple task-oriented assistants, new methods for evaluating such collaborative computer systems are needed. Here, we present a set of criteria for evaluating these systems, calledHallmarks of Human-Machine Collaboration. The Hallmarks build on the success of heuristic evaluation used by the user interface community and past evaluation techniques used in the spoken language and chatbot communities. They consist of observable characteristics indicative of successful collaborative communication, grouped into eight high-level properties: robustness; habitability; mutual contribution of meaningful content; context-awareness; consistent human engagement; provision of rationale; use of elementary concepts to teach and learn new concepts; and successful collaboration. We present examples of how we used these Hallmarks in the DARPA Communicating with Computers (CwC) program to evaluate diverse activities, including story and music generation, interactive building with blocks, and exploration of molecular mechanisms in cancer. We used the Hallmarks as guides for developers and as diagnostics, assessing systems with the Hallmarks to identify strengths and opportunities for improvement using logs from user studies, surveying the human partner, third-party review of creative products, and direct tests. Informal feedback from CwC technology developers indicates that the use of the Hallmarks for program evaluation helped guide development. The Hallmarks also made it possible to identify areas of progress and major gaps in developing systems where the machine is an equal, creative partner.

Light and the circadian clock have a profound effect on the biology of organisms through the regulation of large sets of genes. Toward understanding how light and the circadian clock regulate gene expression, we used genome-wide approaches to identify the direct and indirect targets of the light-responsive and clock-controlled transcription factor ADV-1 in Neurospora crassa. A large proportion of ADV-1 targets were found to be light- and/or clock-controlled, and enriched for genes involved in development, metabolism, cell growth, and cell fusion. We show that ADV-1 is necessary for transducing light and/or temporal information to its immediate downstream targets, including controlling rhythms in genes critical to somatic cell fusion. However, while ADV-1 targets are altered in predictable ways in Δadv-1 cells in response to light, this is not always the case for rhythmic target gene expression. These data suggest that a complex regulatory network downstream of ADV-1 functions to generate distinct temporal dynamics of target gene expression relative to the central clock mechanism.

Bacterial pathogens adapt to changing environments within their hosts, and the signaling molecule adenosine 3', 5'-cyclic monophosphate (cAMP) facilitates this process. In this study, we characterized in vivo DNA binding and gene regulation by the cAMP-responsive protein CRP in M. bovis BCG as a model for tuberculosis (TB)-complex bacteria. Chromatin immunoprecipitation followed by deep-sequencing (ChIP-seq) showed that CRP associates with ∼900 DNA binding regions, most of which occur within genes. The most highly enriched binding region was upstream of a putative copper transporter gene (ctpB), and crp-deleted bacteria showed increased sensitivity to copper toxicity. Detailed mutational analysis of four CRP binding sites upstream of the virulence-associated Rv0249c-Rv0247c succinate dehydrogenase genes demonstrated that CRP directly regulates Rv0249c-Rv0247c expression from two promoters, one of which requires sequences intragenic to Rv0250c for maximum expression. The high percentage of intragenic CRP binding sites and our demonstration that these intragenic DNA sequences significantly contribute to biologically relevant gene expression greatly expand the genome space that must be considered for gene regulatory analyses in mycobacteria. These findings also have practical implications for an important bacterial pathogen in which identification of mutations that affect expression of drug target-related genes is widely used for rapid drug resistance screening.

The filamentous fungus Neurospora crassa responds to light in complex ways. To thoroughly study the transcriptional response of this organism to light, RNA-seq was used to analyze capped and polyadenylated mRNA prepared from mycelium grown for 24 hr in the dark and then exposed to light for 0 (control) 15, 60, 120, and 240 min. More than three-quarters of all defined protein coding genes (79%) were expressed in these cells. The increased sensitivity of RNA-seq compared with previous microarray studies revealed that the RNA levels for 31% of expressed genes were affected two-fold or more by exposure to light. Additionally, a large class of mRNAs, enriched for transcripts specifying products involved in rRNA metabolism, showed decreased expression in response to light, indicating a heretofore undocumented effect of light on this pathway. Based on measured changes in mRNA levels, light generally increases cellular metabolism and at the same time causes significant oxidative stress to the organism. To deal with this stress, protective photopigments are made, antioxidants are produced, and genes involved in ribosome biogenesis are transiently repressed.

We have taken the first steps towards a complete reconstruction of the Mycobacterium tuberculosis regulatory network based on ChIP-Seq and combined this reconstruction with system-wide profiling of messenger RNAs, proteins, metabolites and lipids during hypoxia and re-aeration. Adaptations to hypoxia are thought to have a prominent role in M. tuberculosis pathogenesis. Using ChIP-Seq combined with expression data from the induction of the same factors, we have reconstructed a draft regulatory network based on 50 transcription factors. This network model revealed a direct interconnection between the hypoxic response, lipid catabolism, lipid anabolism and the production of cell wall lipids. As a validation of this model, in response to oxygen availability we observe substantial alterations in lipid content and changes in gene expression and metabolites in corresponding metabolic pathways. The regulatory network reveals transcription factors underlying these changes, allows us to computationally predict expression changes, and indicates that Rv0081 is a regulatory hub.

Background Pathogen metadata includes information about where and when a pathogen was collected and the type of environment it came from. Along with genomic nucleotide sequence data, this metadata is growing rapidly and becoming a valuable resource not only for research but for biosurveillance and public health. However, current freely available tools for analyzing this data are geared towards bioinformaticians and/or do not provide summaries and visualizations needed to readily interpret results. Results We designed a platform to easily access and summarize data about pathogen samples. The software includes a PostgreSQL database that captures metadata useful for disease outbreak investigations, and scripts for downloading and parsing data from NCBI BioSample and BioProject into the database. The software provides a user interface to query metadata and obtain standardized results in an exportable, tab-delimited format. To visually summarize results, the user interface provides a 2D histogram for user-selected metadata types and mapping of geolocated entries. The software is built on the LabKey data platform, an open-source data management platform, which enables developers to add functionalities. We demonstrate the use of the software in querying for a pathogen serovar and for genome sequence identifiers. Conclusions This software enables users to create a local database for pathogen metadata, populate it with data from NCBI, easily query the data, and obtain visual summaries. Some of the components, such as the database, are modular and can be incorporated into other data platforms. The source code is freely available for download at https://github.com/wchangmitre/bioattribution. Electronic supplementary material The online version of this article (doi:10.1186/s12859-016-1231-2) contains supplementary material, which is available to authorized users.

Expression of SigH, one of 12 Mycobacterium tuberculosis alternative sigma factors, is induced by heat, oxidative and nitric oxide stresses. SigH activation has been shown to increase expression of several genes, including genes involved in maintaining redox equilibrium and in protein degradation. However, few of these are known to be directly regulated by SigH. The goal of this project is to comprehensively define the Mycobacterium tuberculosis genes and operons that are directly controlled by SigH in order to gain insight into the role of SigH in regulating M. tuberculosis physiology. We used ChIP-Seq to identify in vivo SigH binding sites throughout the M. tuberculosis genome, followed by quantification of SigH-dependent expression of genes linked to these sites and identification of SigH-regulated promoters. We identified 69 SigH binding sites, which are located both in intergenic regions and within annotated coding sequences in the annotated M. tuberculosis genome. 41 binding sites were linked to genes that showed greater expression following heat stress in a SigH-dependent manner. We identified several genes not previously known to be regulated by SigH, including genes involved in DNA repair, cysteine biosynthesis, translation, and genes of unknown function. Experimental and computational analysis of SigH-regulated promoter sequences within these binding sites identified strong consensus -35 and -10 promoter sequences, but with tolerance for non-consensus bases at specific positions. This comprehensive identification and validation of SigH-regulated genes demonstrates an extended SigH regulon that controls an unexpectedly broad range of stress response functions.

Mycobacterium tuberculosis (Mtb) Cmr (Rv1675c) is a CRP/FNR family transcription factor known to be responsive to cAMP levels and during macrophage infections. However, Cmr's DNA binding properties, cellular targets and overall role in tuberculosis (TB) complex bacteria have not been characterized. In this study, we used experimental and computational approaches to characterize Cmr's DNA binding properties and identify a putative regulon. Cmr binds a 16-bp palindromic site that includes four highly conserved nucleotides that are required for DNA binding. A total of 368 binding sites, distributed in clusters among ~200 binding regions throughout the Mycobacterium bovis BCG genome, were identified using ChIP-seq. One of the most enriched Cmr binding sites was located upstream of the cmr promoter, and we demonstrated that expression of cmr is autoregulated. cAMP affected Cmr binding at a subset of DNA loci in vivo and in vitro, including multiple sites adjacent to members of the DosR (DevR) dormancy regulon. Our findings of cooperative binding of Cmr to these DNA regions and the regulation by Cmr of the DosR-regulated virulence gene Rv2623 demonstrate the complexity of Cmr-mediated gene regulation and suggest a role for Cmr in the biology of persistent TB infection.

Transcription factors (TFs) play a central role in regulating gene expression in all bacteria. Yet until recently, studies of TF binding were limited to a small number of factors at a few genomic locations. Chromatin immunoprecipitation followed by sequencing (ChIP-Seq) provides the ability to map binding sites globally for TFs, and the scalability of the technology enables the ability to map binding sites for every DNA binding protein in a prokaryotic organism. We have developed a protocol for ChIP-Seq tailored for use with mycobacteria and an analysis pipeline for processing the resulting data. The protocol and pipeline have been used to map over 100 TFs from Mycobacterium tuberculosis , as well as numerous TFs from related mycobacteria and other bacteria. The resulting data provide evidence that the long-accepted spatial relationship between TF binding site, promoter motif, and the corresponding regulated gene may be too simple a paradigm, failing to adequately capture the variety of TF binding sites found in prokaryotes. In this article we describe the protocol and analysis pipeline, the validation of these methods, and the results of applying these methods to M. tuberculosis .

Background Current biosensors are designed to target and react to specific nucleic acid sequences or structural epitopes. These 'target-specific' platforms require creation of new physical capture reagents when new organisms are targeted. An 'open-target' approach to DNA microarray biosensing is proposed and substantiated using laboratory generated data. The microarray consisted of 12,900 25 bp oligonucleotide capture probes derived from a statistical model trained on randomly selected genomic segments of pathogenic prokaryotic organisms. Open-target detection of organisms was accomplished using a reference library of hybridization patterns for three test organisms whose DNA sequences were not included in the design of the microarray probes. Results A multivariate mathematical model based on the partial least squares regression (PLSR) was developed to detect the presence of three test organisms in mixed samples. When all 12,900 probes were used, the model correctly detected the signature of three test organisms in all mixed samples (mean(R2)) = 0.76, CI = 0.95), with a 6% false positive rate. A sampling algorithm was then developed to sparsely sample the probe space for a minimal number of probes required to capture the hybridization imprints of the test organisms. The PLSR detection model was capable of correctly identifying the presence of the three test organisms in all mixed samples using only 47 probes (mean(R2)) = 0.77, CI = 0.95) with nearly 100% specificity. Conclusions We conceived an 'open-target' approach to biosensing, and hypothesized that a relatively small, non-specifically designed, DNA microarray is capable of identifying the presence of multiple organisms in mixed samples. Coupled with a mathematical model applied to laboratory generated data, and sparse sampling of capture probes, the prototype microarray platform was able to capture the signature of each organism in all mixed samples with high sensitivity and specificity. It was demonstrated that this new approach to biosensing closely follows the principles of sparse sensing., Mitre Corporation

Human Serum paraoxonase 1 (HuPON1) is an enzyme that has been shown to hydrolyze a variety of chemicals including the nerve agent VX. While wildtype HuPON1 does not exhibit sufficient activity against VX to be used as an in vivo countermeasure, it has been suggested that increasing HuPON1's organophosphorous hydrolase activity by one or two orders of magnitude would make the enzyme suitable for this purpose. The binding interaction between HuPON1 and VX has recently been modeled, but the mechanism for VX hydrolysis is still unknown. In this study, we created a transition state model for VX hydrolysis (VXts) in water using quantum mechanical/molecular mechanical simulations, and docked the transition state model to 22 experimentally characterized HuPON1 variants using AutoDock Vina. The HuPON1-VXts complexes were grouped by reaction mechanism using a novel clustering procedure. The average Vina interaction energies for different clusters were compared to the experimentally determined activities of HuPON1 variants to determine which computational procedures best predict how well HuPON1 variants will hydrolyze VX. The analysis showed that only conformations which have the attacking hydroxyl group of VXts coordinated by the sidechain oxygen of D269 have a significant correlation with experimental results. The results from this study can be used for further characterization of how HuPON1 hydrolyzes VX and design of HuPON1 variants with increased activity against VX., United States. Defense Threat Reduction Agency

The concept of co-design is common in engineering, where it is necessary, for example, to determine the optimal partitioning between hardware and software of the implementation of a system features. Here we propose to adapt co-design methodologies for synthetic biology. As a test case, we have designed an environmental sensing device that detects the presence of three chemicals, and returns an output only if at least two of the three chemicals are present. We show that the logical operations can be implemented in three different design domains: (1) the transcriptional domain using synthetically designed hybrid promoters, (2) the protein domain using bi-molecular fluorescence complementation, and (3) the fluorescence domain using spectral unmixing and relying on electronic processing. We discuss how these heterogeneous design strategies could be formalized to develop co-design algorithms capable of identifying optimal designs meeting user specifications.

Human serum paraoxonase 1 (HuPON1) is an enzyme that can hydrolyze various chemical warfare nerve agents including VX. A previous study has suggested that increasing HuPON1’s VX hydrolysis activity one to two orders of magnitude would make the enzyme an effective countermeasure for in vivo use against VX. This study helps facilitate further engineering of HuPON1 for enhanced VX-hydrolase activity by computationally characterizing HuPON1’s tertiary structure and how HuPON1 binds VX. HuPON1’s structure is first predicted through two homology modeling procedures. Docking is then performed using four separate methods, and the stability of each bound conformation is analyzed through molecular dynamics and solvated interaction energy calculations. The results show that VX’s lone oxygen atom has a strong preference for forming a direct electrostatic interaction with HuPON1’s active site calcium ion. Various HuPON1 residues are also detected that are in close proximity to VX and are therefore potential targets for future mutagenesis studies. These include E53, H115, N168, F222, N224, L240, D269, I291, F292, and V346. Additionally, D183 was found to have a predicted pKa near physiological pH. Given D183’s location in HuPON1’s active site, this residue could potentially act as a proton donor or accepter during hydrolysis. The results from the binding simulations also indicate that steered molecular dynamics can potentially be used to obtain accurate binding predictions even when starting with a closed conformation of a protein’s binding or active site.

Protein-protein docking is the computational prediction of protein complex structure given the individually solved component protein structures. It is an important means for understanding the physicochemical forces that underlie macromolecular interactions and a valuable tool for modeling protein complex structures. Here, we report an overview of protein-protein docking with specific emphasis on our Fast Fourier Transform-based rigid-body docking program ZDOCK, which is consistently rated as one of the most accurate docking programs in the Critical Assessment of Predicted Interactions (CAPRI), a series of community-wide blind tests. We also investigate ZDOCK's performance on a non-redundant protein complex benchmark. Finally, we perform regression analysis to better understand the strengths and weaknesses of ZDOCK and to suggest areas of future development for protein-docking algorithms in general.

Protein-protein docking is the computational prediction of protein complex structure given the individually solved component protein structures. It is an important means for understanding the physicochemical forces that underlie macromolecular interactions and a valuable tool for modeling protein complex structures. Here, we report an overview of protein-protein docking with specific emphasis on our Fast Fourier Transform-based rigid-body docking program ZDOCK, which is consistently rated as one of the most accurate docking programs in the Critical Assessment of Predicted Interactions (CAPRI), a series of community-wide blind tests. We also investigate ZDOCK's performance on a non-redundant protein complex benchmark. Finally, we perform regression analysis to better understand the strengths and weaknesses of ZDOCK and to suggest areas of future development for protein-docking algorithms in general.

Museum fish collections possess a wealth of anatomical and morphological data that are essential for documenting and understanding biodiversity. Obtaining access to specimens for research, however, is not always practical and frequently conflicts with the need to maintain the physical integrity of specimens and the collection as a whole. Non-invasive three-dimensional (3D) digital imaging therefore serves a critical role in facilitating the digitization of these specimens for anatomical and morphological analysis as well as facilitating an efficient method for online storage and sharing of this imaging data. Here we describe the development of the Digital Fish Library (DFL, http://www.digitalfishlibrary.org), an online digital archive of high-resolution, high-contrast, magnetic resonance imaging (MRI) scans of the soft tissue anatomy of an array of fishes preserved in the Marine Vertebrate Collection of Scripps Institution of Oceanography. We have imaged and uploaded MRI data for over 300 marine and freshwater species, developed a data archival and retrieval system with a web-based image analysis and visualization tool, and integrated these into the public DFL website to disseminate data and associated metadata freely over the web. We show that MRI is a rapid and powerful method for accurately depicting the in-situ soft-tissue anatomy of preserved fishes in sufficient detail for large-scale comparative digital morphology. However these 3D volumetric data require a sophisticated computational and archival infrastructure in order to be broadly accessible to researchers and educators.

Background Phylogenetic trees are widely used to visualize evolutionary relationships between different organisms or samples of the same organism. There exists a variety of both free and commercial tree visualization software available, but limitations in these programs often require researchers to use multiple programs for analysis, annotation, and the production of publication-ready images. Results We present TreeViewJ, a Java tool for visualizing, editing and analyzing phylogenetic trees. The software allows researchers to color and change the width of branches that they wish to highlight, and add names to nodes. If collection dates are available for taxa, the software can map them onto a timeline, and sort the tree in ascending or descending date order. Conclusion TreeViewJ is a tool for researchers to visualize, edit, "decorate," and produce publication-ready images of phylogenetic trees. It is open-source, and released under an GPL license, and available at http://treeviewj.sourceforge.net.

Background Current sequencing technology makes it practical to sequence many samples of a given organism, raising new challenges for the processing and interpretation of large genomics data sets with associated metadata. Traditional computational phylogenetic methods are ideal for studying the evolution of gene/protein families and using those to infer the evolution of an organism, but are less than ideal for the study of the whole organism mainly due to the presence of insertions/deletions/rearrangements. These methods provide the researcher with the ability to group a set of samples into distinct genotypic groups based on sequence similarity, which can then be associated with metadata, such as host information, pathogenicity, and time or location of occurrence. Genotyping is critical to understanding, at a genomic level, the origin and spread of infectious diseases. Increasingly, genotyping is coming into use for disease surveillance activities, as well as for microbial forensics. The classic genotyping approach has been based on phylogenetic analysis, starting with a multiple sequence alignment. Genotypes are then established by expert examination of phylogenetic trees. However, these traditional single-processor methods are suboptimal for rapidly growing sequence datasets being generated by next-generation DNA sequencing machines, because they increase in computational complexity quickly with the number of sequences. Results Nephele is a suite of tools that uses the complete composition vector algorithm to represent each sequence in the dataset as a vector derived from its constituent k-mers by passing the need for multiple sequence alignment, and affinity propagation clustering to group the sequences into genotypes based on a distance measure over the vectors. Our methods produce results that correlate well with expert-defined clades or genotypes, at a fraction of the computational cost of traditional phylogenetic methods run on traditional hardware. Nephele can use the open-source Hadoop implementation of MapReduce to parallelize execution using multiple compute nodes. We were able to generate a neighbour-joined tree of over 10,000 16S samples in less than 2 hours. Conclusions We conclude that using Nephele can substantially decrease the processing time required for generating genotype trees of tens to hundreds of organisms at genome scale sequence coverage.