Dual-Force ISOMAP: A New Relevance Feedback Method for Medical Image Retrieval

With great potential for assisting radiological image interpretation and decision making, content-based image retrieval in the medical domain has become a hot topic in recent years. Many methods to enhance the performance of content-based medical image retrieval have been proposed, among which the relevance feedback (RF) scheme is one of the most promising. Given user feedback information, RF algorithms interactively learn a user’s preferences to bridge the “semantic gap” between low-level computerized visual features and high-level human semantic perception and thus improve retrieval performance. However, most existing RF algorithms perform in the original high-dimensional feature space and ignore the manifold structure of the low-level visual features of images. In this paper, we propose a new method, termed dual-force ISOMAP (DFISOMAP), for content-based medical image retrieval. Under the assumption that medical images lie on a low-dimensional manifold embedded in a high-dimensional ambient space, DFISOMAP operates in the following three stages. First, the geometric structure of positive examples in the learned low-dimensional embedding is preserved according to the isometric feature mapping (ISOMAP) criterion. To precisely model the geometric structure, a reconstruction error constraint is also added. Second, the average distance between positive and negative examples is maximized to separate them; this margin maximization acts as a force that pushes negative examples far away from positive examples. Finally, the similarity propagation technique is utilized to provide negative examples with another force that will pull them back into the negative sample set. We evaluate the proposed method on a subset of the IRMA medical image dataset with a RF-based medical image retrieval framework. Experimental results show that DFISOMAP outperforms popular approaches for content-based medical image retrieval in terms of accuracy and stability.     

Fuzzy Nonlinear Proximal Support Vector Machine for Land Extraction Based on Remote Sensing Image

Currently, remote sensing technologies were widely employed in the dynamic monitoring of the land. This paper presented an algorithm named fuzzy nonlinear proximal support vector machine (FNPSVM) by basing on ETM⁺ remote sensing image. This algorithm is applied to extract various types of lands of the city Da’an in northern China. Two multi-category strategies, namely “one-against-one” and “one-against-rest” for this algorithm were described in detail and then compared. A fuzzy membership function was presented to reduce the effects of noises or outliers on the data samples. The approaches of feature extraction, feature selection, and several key parameter settings were also given. Numerous experiments were carried out to evaluate its performances including various accuracies (overall accuracies and kappa coefficient), stability, training speed, and classification speed. The FNPSVM classifier was compared to the other three classifiers including the maximum likelihood classifier (MLC), back propagation neural network (BPN), and the proximal support vector machine (PSVM) under different training conditions. The impacts of the selection of training samples, testing samples and features on the four classifiers were also evaluated in these experiments.     

Word Decoding of Protein Amino Acid Sequences with Availability Analysis: A Linguistic Approach

The amino acid sequences of proteins determine their three-dimensional structures and functions. However, how sequence information is related to structures and functions is still enigmatic. In this study, we show that at least a part of the sequence information can be extracted by treating amino acid sequences of proteins as a collection of English words, based on a working hypothesis that amino acid sequences of proteins are composed of short constituent amino acid sequences (SCSs) or “words”. We first confirmed that the English language highly likely follows Zipf''s law, a special case of power law. We found that the rank-frequency plot of SCSs in proteins exhibits a similar distribution when low-rank tails are excluded. In comparison with natural English and “compressed” English without spaces between words, amino acid sequences of proteins show larger linear ranges and smaller exponents with heavier low-rank tails, demonstrating that the SCS distribution in proteins is largely scale-free. A distribution pattern of SCSs in proteins is similar among species, but species-specific features are also present. Based on the availability scores of SCSs, we found that sequence motifs are enriched in high-availability sites (i.e., “key words”) and vice versa. In fact, the highest availability peak within a given protein sequence often directly corresponds to a sequence motif. The amino acid composition of high-availability sites within motifs is different from that of entire motifs and all protein sequences, suggesting the possible functional importance of specific SCSs and their compositional amino acids within motifs. We anticipate that our availability-based word decoding approach is complementary to sequence alignment approaches in predicting functionally important sites of unknown proteins from their amino acid sequences.     

A combination dimensionality reduction approach to codon position patterns of eubacteria based on their complete genomes

Graphical techniques have become powerful tools for the visualization and analysis of complicated biological systems. However, we cannot give such a graphical representation in a 2D/3D space when the dimensions of the represented data are more than three dimensions. The proposed method, a combination dimensionality reduction approach (CDR), consists of two parts: (i) principal component analysis (PCA) with a newly defined parameter ρ and (ii) locally linear embedding (LLE) with a proposed graphical selection for its optional parameter k. The CDR approach with ρ and k not only avoids loss of principal information, but also sufficiently well preserves the global high-dimensional structures in low-dimensional space such as 2D or 3D. The applications of the CDR on characteristic analysis at different codon positions in genome show that the method is a useful tool by which biologists could find useful biological knowledge.     

Coupled Human-Environment Dynamics of Forest Pest Spread and Control in a Multi-Patch,Stochastic Setting

Background

The transportation of camp firewood infested by non-native forest pests such as Asian long-horned beetle (ALB) and emerald ash borer (EAB) has severe impacts on North American forests. Once invasive forest pests are established, it can be difficult to eradicate them. Hence, preventing the long-distance transport of firewood by individuals is crucial.

Methods

Here we develop a stochastic simulation model that captures the interaction between forest pest infestations and human decisions regarding firewood transportation. The population of trees is distributed across 10 patches (parks) comprising a “low volume” partition of 5 patches that experience a low volume of park visitors, and a “high volume” partition of 5 patches experiencing a high visitor volume. The infestation spreads within a patch—and also between patches—according to the probability of between-patch firewood transportation. Individuals decide to transport firewood or buy it locally based on the costs of locally purchased versus transported firewood, social norms, social learning, and level of concern for observed infestations.

Results

We find that the average time until a patch becomes infested depends nonlinearly on many model parameters. In particular, modest increases in the tree removal rate, modest increases in public concern for infestation, and modest decreases in the cost of locally purchased firewood, relative to baseline (current) values, cause very large increases in the average time until a patch becomes infested due to firewood transport from other patches, thereby better preventing long-distance spread. Patches that experience lower visitor volumes benefit more from firewood movement restrictions than patches that experience higher visitor volumes. Also, cross–patch infestations not only seed new infestations, they can also worsen existing infestations to a surprising extent: long-term infestations are more intense in the high volume patches than the low volume patches, even when infestation is already endemic everywhere.

Conclusions

The success of efforts to prevent long-distance spread of forest pests may depend sensitively on the interaction between outbreak dynamics and human social processes, with similar levels of effort producing very different outcomes depending on where the coupled human and natural system exists in parameter space. Further development of such modeling approaches through better empirical validation should yield more precise recommendations for ways to optimally prevent the long-distance spread of invasive forest pests.     

Affective Video Retrieval: Violence Detection in Hollywood Movies by Large-Scale Segmental Feature Extraction

Without doubt general video and sound, as found in large multimedia archives, carry emotional information. Thus, audio and video retrieval by certain emotional categories or dimensions could play a central role for tomorrow''s intelligent systems, enabling search for movies with a particular mood, computer aided scene and sound design in order to elicit certain emotions in the audience, etc. Yet, the lion''s share of research in affective computing is exclusively focusing on signals conveyed by humans, such as affective speech. Uniting the fields of multimedia retrieval and affective computing is believed to lend to a multiplicity of interesting retrieval applications, and at the same time to benefit affective computing research, by moving its methodology “out of the lab” to real-world, diverse data. In this contribution, we address the problem of finding “disturbing” scenes in movies, a scenario that is highly relevant for computer-aided parental guidance. We apply large-scale segmental feature extraction combined with audio-visual classification to the particular task of detecting violence. Our system performs fully data-driven analysis including automatic segmentation. We evaluate the system in terms of mean average precision (MAP) on the official data set of the MediaEval 2012 evaluation campaign''s Affect Task, which consists of 18 original Hollywood movies, achieving up to .398 MAP on unseen test data in full realism. An in-depth analysis of the worth of individual features with respect to the target class and the system errors is carried out and reveals the importance of peak-related audio feature extraction and low-level histogram-based video analysis.     

Functional Brain Connectivity as a New Feature for P300 Speller

The brain is a large-scale complex network often referred to as the “connectome”. Cognitive functions and information processing are mainly based on the interactions between distant brain regions. However, most of the ‘feature extraction’ methods used in the context of Brain Computer Interface (BCI) ignored the possible functional relationships between different signals recorded from distinct brain areas. In this paper, the functional connectivity quantified by the phase locking value (PLV) was introduced to characterize the evoked responses (ERPs) obtained in the case of target and non-targets visual stimuli. We also tested the possibility of using the functional connectivity in the context of ‘P300 speller’. The proposed approach was compared to the well-known methods proposed in the state of the art of “P300 Speller”, mainly the peak picking, the area, time/frequency based features, the xDAWN spatial filtering and the stepwise linear discriminant analysis (SWLDA). The electroencephalographic (EEG) signals recorded from ten subjects were analyzed offline. The results indicated that phase synchrony offers relevant information for the classification in a P300 speller. High synchronization between the brain regions was clearly observed during target trials, although no significant synchronization was detected for a non-target trial. The results showed also that phase synchrony provides higher performance than some existing methods for letter classification in a P300 speller principally when large number of trials is available. Finally, we tested the possible combination of both approaches (classical features and phase synchrony). Our findings showed an overall improvement of the performance of the P300-speller when using Peak picking, the area and frequency based features. Similar performances were obtained compared to xDAWN and SWLDA when using large number of trials.     

Protein functional features are reflected in the patterns of mRNA translation speed

Background

The degeneracy of the genetic code makes it possible for the same amino acid string to be coded by different messenger RNA (mRNA) sequences. These “synonymous mRNAs” may differ largely in a number of aspects related to their overall translational efficiency, such as secondary structure content and availability of the encoded transfer RNAs (tRNAs). Consequently, they may render different yields of the translated polypeptides. These mRNA features related to translation efficiency are also playing a role locally, resulting in a non-uniform translation speed along the mRNA, which has been previously related to some protein structural features and also used to explain some dramatic effects of “silent” single-nucleotide-polymorphisms (SNPs). In this work we perform the first large scale analysis of the relationship between three experimental proxies of mRNA local translation efficiency and the local features of the corresponding encoded proteins.

Results

We found that a number of protein functional and structural features are reflected in the patterns of ribosome occupancy, secondary structure and tRNA availability along the mRNA. One or more of these proxies of translation speed have distinctive patterns around the mRNA regions coding for certain protein local features. In some cases the three patterns follow a similar trend. We also show specific examples where these patterns of translation speed point to the protein’s important structural and functional features.

Conclusions

This support the idea that the genome not only codes the protein functional features as sequences of amino acids, but also as subtle patterns of mRNA properties which, probably through local effects on the translation speed, have some consequence on the final polypeptide. These results open the possibility of predicting a protein’s functional regions based on a single genomic sequence, and have implications for heterologous protein expression and fine-tuning protein function.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-015-1734-7) contains supplementary material, which is available to authorized users.     

The Touch and Zap Method for In Vivo Whole-Cell Patch Recording of Intrinsic and Visual Responses of Cortical Neurons and Glial Cells

Whole-cell patch recording is an essential tool for quantitatively establishing the biophysics of brain function, particularly in vivo. This method is of particular interest for studying the functional roles of cortical glial cells in the intact brain, which cannot be assessed with extracellular recordings. Nevertheless, a reasonable success rate remains a challenge because of stability, recording duration and electrical quality constraints, particularly for voltage clamp, dynamic clamp or conductance measurements. To address this, we describe “Touch and Zap”, an alternative method for whole-cell patch clamp recordings, with the goal of being simpler, quicker and more gentle to brain tissue than previous approaches. Under current clamp mode with a continuous train of hyperpolarizing current pulses, seal formation is initiated immediately upon cell contact, thus the “Touch”. By maintaining the current injection, whole-cell access is spontaneously achieved within seconds from the cell-attached configuration by a self-limited membrane electroporation, or “Zap”, as seal resistance increases. We present examples of intrinsic and visual responses of neurons and putative glial cells obtained with the revised method from cat and rat cortices in vivo. Recording parameters and biophysical properties obtained with the Touch and Zap method compare favourably with those obtained with the traditional blind patch approach, demonstrating that the revised approach does not compromise the recorded cell. We find that the method is particularly well-suited for whole-cell patch recordings of cortical glial cells in vivo, targeting a wider population of this cell type than the standard method, with better access resistance. Overall, the gentler Touch and Zap method is promising for studying quantitative functional properties in the intact brain with minimal perturbation of the cell''s intrinsic properties and local network. Because the Touch and Zap method is performed semi-automatically, this approach is more reproducible and less dependent on experimenter technique.     

Manifold learning for human population structure studies

The dimension of the population genetics data produced by next-generation sequencing platforms is extremely high. However, the "intrinsic dimensionality" of sequence data, which determines the structure of populations, is much lower. This motivates us to use locally linear embedding (LLE) which projects high dimensional genomic data into low dimensional, neighborhood preserving embedding, as a general framework for population structure and historical inference. To facilitate application of the LLE to population genetic analysis, we systematically investigate several important properties of the LLE and reveal the connection between the LLE and principal component analysis (PCA). Identifying a set of markers and genomic regions which could be used for population structure analysis will provide invaluable information for population genetics and association studies. In addition to identifying the LLE-correlated or PCA-correlated structure informative marker, we have developed a new statistic that integrates genomic information content in a genomic region for collectively studying its association with the population structure and LASSO algorithm to search such regions across the genomes. We applied the developed methodologies to a low coverage pilot dataset in the 1000 Genomes Project and a PHASE III Mexico dataset of the HapMap. We observed that 25.1%, 44.9% and 21.4% of the common variants and 89.2%, 92.4% and 75.1% of the rare variants were the LLE-correlated markers in CEU, YRI and ASI, respectively. This showed that rare variants, which are often private to specific populations, have much higher power to identify population substructure than common variants. The preliminary results demonstrated that next generation sequencing offers a rich resources and LLE provide a powerful tool for population structure analysis.     

Live cell imaging of the assembly, disassembly, and actin cable-dependent movement of endosomes and actin patches in the budding yeast, Saccharomyces cerevisiae

Using FM4-64 to label endosomes and Abp1p-GFP or Sac6p-GFP to label actin patches, we find that (1) endosomes colocalize with actin patches as they assemble at the bud cortex; (2) endosomes colocalize with actin patches as they undergo linear, retrograde movement from buds toward mother cells; and (3) actin patches interact with and disassemble at FM4-64–labeled internal compartments. We also show that retrograde flow of actin cables mediates retrograde actin patch movement. An Arp2/3 complex mutation decreases the frequency of cortical, nonlinear actin patch movements, but has no effect on the velocity of linear, retrograde actin patch movement. Rather, linear actin patch movement occurs at the same velocity and direction as the movement of actin cables. Moreover, actin patches require actin cables for retrograde movements and colocalize with actin cables as they undergo retrograde movement. Our studies support a mechanism whereby actin cables serve as “conveyor belts” for retrograde movement and delivery of actin patches/endosomes to FM4-64–labeled internal compartments.     

The Medical Practitioner,Alcoholism and Motivation

Adverse influences on motivation for recovery from alcoholism must be searched for in three areas: society, the medical practitioner and the patient. Society is ambivalent because there is a vicarious release through identification with the cheerful “drunk” coupled with unconscious envy and resentment leading to punitive action.The current “alcohol culture” decrees that to drink is to be well, not to drink is to be ill.The medical profession attempts to suppress, deny, rationalize or reject the problem of alcoholism because it involves a change in attitude and recognition of limitations.The alcoholic patient has a notorious lack of motivation, but this must be recognized as a symptom of his disease, and with certain techniques this symptom is treatable. Furthermore, motivation fluctuates and many opportunities for treatment are available when the medical practitioner can detect that motivation is high. At times a coercive approach is required, at times a permissive one; and the optimal use of such approaches will increase the motivation to an effective level.     

Basis for Cumulative Cultural Evolution in Chimpanzees: Social Learning of a More Efficient Tool-Use Technique

Background

The evidence for culture in non-human animals has been growing incrementally over the past two decades. However, the ability for cumulative cultural evolution, with successive generations building on earlier achievements, in non-human animals remains debated. Faithful social learning of incremental improvements in technique is considered to be a defining feature of human culture, differentiating human from non-human cultures. This study presents the first experimental evidence for chimpanzees'' social transmission of a more efficient tool-use technique invented by a conspecific group member.

Methodology/Principal Findings

The chimpanzees were provided with a straw-tube, and spontaneously demonstrated two different techniques in obtaining juice through a small hole: “dipping” and “straw-sucking”. Both the “dipping” and “straw-sucking” techniques depended on the use of the same tool (straw-tube) for the same target (juice) accessible from exactly the same location (small hole 1 cm in diameter). Therefore the difference between “dipping” and “straw-sucking” was only in “technique”. Although the two techniques differed significantly in their efficiency, their cognitive and perceptuo-motor complexity were comparable. All five chimpanzees who initially performed the “dipping” technique switched to using the more efficient “straw-sucking” technique upon observing a conspecific or human demonstrate the more proficient alternate “straw-sucking” technique.

Conclusions/Significance

The social learning mechanism involved here was clearly not local or stimulus enhancement, but imitation or emulation of a tool-use technique. When there is no biologically relevant difference in cognitive or perceptuo-motor complexity between two techniques, and when chimpanzees are dissatisfied with their own technique, chimpanzees may socially learn an improved technique upon close observation of a proficient demonstrator. This study provides important insights into the cognitive basis for cumulative culture in chimpanzees, and also suggests possible conditions in which cumulative cultural evolution could arise even in non-human animals.     

Spatiotemporal Dynamics of Actomyosin Networks

Rhodamine–phalloidin-labeled actin filaments were visualized gliding over a skeletal heavy meromyosin (HMM)-coated surface. Experiments at low filament densities showed that when two filaments collided, their paths were affected in a manner that depended on collision angle. Some collisions resulted in complete alignment of the filament paths; in others, the filaments crossed over one another. Filament crossover or alignment was equally probable at ∼40° contact angle. Filaments often underwent significant bending during collision and analysis of filament shape indicated an energy requirement of ∼13 k_BT. Experiments were performed over a wide range of HMM surface density and actin filament bulk concentration. Actin filament gliding speed and path persistence plateaued above a critical HMM surface density, and at high (micromolar) actin filament concentrations, filament motion became dramatically aligned in a common direction. Spatiotemporal features of alignment behavior were determined by correlation analysis, supported by simulations. The thermal drift of individual filament tracks was suppressed as the population became more oriented. Spatial correlation analysis revealed that long-range alignment was due to incremental recruitment rather than fusion of locally ordered seed domains. The global alignment of filament movement, described by an “order parameter,” peaked at optimal actin concentrations and myosin surface densities, in contrast to previous predictions of a critical phase transition. Either hydrodynamic coupling or exchange of filaments between the surface bound and adjacent bulk phase layers might degrade order at high actin filament concentration, and high HMM surface densities might decrease alignment probability during collisions. Our results are compatible with generation of long-range order from mechanical interaction between individual actin filaments. Furthermore, we show that randomly oriented myosin motors align relatively short, submicrometer actin filaments into motile surface domains that extend over many tens of micrometers and these patterns persist for several minutes.     

The Route to an Integrative Associative Memory Is Influenced by Emotion

Though the hippocampus typically has been implicated in processes related to associative binding, special types of associations – such as those created by integrative mental imagery – may be supported by processes implemented in other medial temporal-lobe or sensory processing regions. Here, we investigated what neural mechanisms underlie the formation and subsequent retrieval of integrated mental images, and whether those mechanisms differ based on the emotionality of the integration (i.e., whether it contains an emotional item or not). Participants viewed pairs of words while undergoing a functional MRI scan. They were instructed to imagine the two items separately from one another (“non-integrative” study) or as a single, integrated mental image (“integrative” study). They provided ratings of how successful they were at generating vivid images that fit the instructions. They were then given a surprise associative recognition test, also while undergoing an fMRI scan. The cuneus showed parametric correspondence to increasing imagery success selectively during encoding and retrieval of emotional integrations, while the parahippocampal gyri and prefrontal cortices showed parametric correspondence during the encoding and retrieval of non-emotional integrations. Connectivity analysis revealed that selectively during negative integration, left amygdala activity was negatively correlated with frontal and hippocampal activity. These data indicate that individuals utilize two different neural routes for forming and retrieving integrations depending on their emotional content, and they suggest a potentially disruptive role for the amygdala on frontal and medial-temporal regions during negative integration.     

Incorporating homologues into sequence embeddings for protein analysis

Statistical and learning techniques are becoming increasingly popular for different tasks in bioinformatics. Many of the most powerful statistical and learning techniques are applicable to points in a Euclidean space but not directly applicable to discrete sequences such as protein sequences. One way to apply these techniques to protein sequences is to embed the sequences into a Euclidean space and then apply these techniques to the embedded points. In this work we introduce a biologically motivated sequence embedding, the homology kernel, which takes into account intuitions from local alignment, sequence homology, and predicted secondary structure. This embedding allows us to directly apply learning techniques to protein sequences. We apply the homology kernel in several ways. We demonstrate how the homology kernel can be used for protein family classification and outperforms state-of-the-art methods for remote homology detection. We show that the homology kernel can be used for secondary structure prediction and is competitive with popular secondary structure prediction methods. Finally, we show how the homology kernel can be used to incorporate information from homologous sequences in local sequence alignment.