Optimized approach to decision fusion of heterogeneous data for breast cancer diagnosis.

As more diagnostic testing options become available to physicians, it becomes more difficult to combine various types of medical information together in order to optimize the overall diagnosis. To improve diagnostic performance, here we introduce an approach to optimize a decision-fusion technique to combine heterogeneous information, such as from different modalities, feature categories, or institutions. For classifier comparison we used two performance metrics: The receiving operator characteristic (ROC) area under the curve [area under the ROC curve (AUC)] and the normalized partial area under the curve (pAUC). This study used four classifiers: Linear discriminant analysis (LDA), artificial neural network (ANN), and two variants of our decision-fusion technique, AUC-optimized (DF-A) and pAUC-optimized (DF-P) decision fusion. We applied each of these classifiers with 100-fold cross-validation to two heterogeneous breast cancer data sets: One of mass lesion features and a much more challenging one of microcalcification lesion features. For the calcification data set, DF-A outperformed the other classifiers in terms of AUC (p < 0.02) and achieved AUC=0.85 +/- 0.01. The DF-P surpassed the other classifiers in terms of pAUC (p < 0.01) and reached pAUC=0.38 +/- 0.02. For the mass data set, DF-A outperformed both the ANN and the LDA (p < 0.04) and achieved AUC=0.94 +/- 0.01. Although for this data set there were no statistically significant differences among the classifiers' pAUC values (pAUC=0.57 +/- 0.07 to 0.67 +/- 0.05, p > 0.10), the DF-P did significantly improve specificity versus the LDA at both 98% and 100% sensitivity (p < 0.04). In conclusion, decision fusion directly optimized clinically significant performance measures, such as AUC and pAUC, and sometimes outperformed two well-known machine-learning techniques when applied to two different breast cancer data sets.

An active learning approach for rapid characterization of endothelial cells in human tumors.

Currently, no available pathological or molecular measures of tumor angiogenesis predict response to antiangiogenic therapies used in clinical practice. Recognizing that tumor endothelial cells (EC) and EC activation and survival signaling are the direct targets of these therapies, we sought to develop an automated platform for quantifying activity of critical signaling pathways and other biological events in EC of patient tumors by histopathology. Computer image analysis of EC in highly heterogeneous human tumors by a statistical classifier trained using examples selected by human experts performed poorly due to subjectivity and selection bias. We hypothesized that the analysis can be optimized by a more active process to aid experts in identifying informative training examples. To test this hypothesis, we incorporated a novel active learning (AL) algorithm into FARSIGHT image analysis software that aids the expert by seeking out informative examples for the operator to label. The resulting FARSIGHT-AL system identified EC with specificity and sensitivity consistently greater than 0.9 and outperformed traditional supervised classification algorithms. The system modeled individual operator preferences and generated reproducible results. Using the results of EC classification, we also quantified proliferation (Ki67) and activity in important signal transduction pathways (MAP kinase, STAT3) in immunostained human clear cell renal cell carcinoma and other tumors. FARSIGHT-AL enables characterization of EC in conventionally preserved human tumors in a more automated process suitable for testing and validating in clinical trials. The results of our study support a unique opportunity for quantifying angiogenesis in a manner that can now be tested for its ability to identify novel predictive and response biomarkers.

Gene selection using iterative feature elimination random forests for survival outcomes.

Although many feature selection methods for classification have been developed, there is a need to identify genes in high-dimensional data with censored survival outcomes. Traditional methods for gene selection in classification problems have several drawbacks. First, the majority of the gene selection approaches for classification are single-gene based. Second, many of the gene selection procedures are not embedded within the algorithm itself. The technique of random forests has been found to perform well in high-dimensional data settings with survival outcomes. It also has an embedded feature to identify variables of importance. Therefore, it is an ideal candidate for gene selection in high-dimensional data with survival outcomes. In this paper, we develop a novel method based on the random forests to identify a set of prognostic genes. We compare our method with several machine learning methods and various node split criteria using several real data sets. Our method performed well in both simulations and real data analysis.Additionally, we have shown the advantages of our approach over single-gene-based approaches. Our method incorporates multivariate correlations in microarray data for survival outcomes. The described method allows us to better utilize the information available from microarray data with survival outcomes.

Genetic signatures in the envelope glycoproteins of HIV-1 that associate with broadly neutralizing antibodies.

A steady increase in knowledge of the molecular and antigenic structure of the gp120 and gp41 HIV-1 envelope glycoproteins (Env) is yielding important new insights for vaccine design, but it has been difficult to translate this information to an immunogen that elicits broadly neutralizing antibodies. To help bridge this gap, we used phylogenetically corrected statistical methods to identify amino acid signature patterns in Envs derived from people who have made potently neutralizing antibodies, with the hypothesis that these Envs may share common features that would be useful for incorporation in a vaccine immunogen. Before attempting this, essentially as a control, we explored the utility of our computational methods for defining signatures of complex neutralization phenotypes by analyzing Env sequences from 251 clonal viruses that were differentially sensitive to neutralization by the well-characterized gp120-specific monoclonal antibody, b12. We identified ten b12-neutralization signatures, including seven either in the b12-binding surface of gp120 or in the V2 region of gp120 that have been previously shown to impact b12 sensitivity. A simple algorithm based on the b12 signature pattern was predictive of b12 sensitivity/resistance in an additional blinded panel of 57 viruses. Upon obtaining these reassuring outcomes, we went on to apply these same computational methods to define signature patterns in Env from HIV-1 infected individuals who had potent, broadly neutralizing responses. We analyzed a checkerboard-style neutralization dataset with sera from 69 HIV-1-infected individuals tested against a panel of 25 different Envs. Distinct clusters of sera with high and low neutralization potencies were identified. Six signature positions in Env sequences obtained from the 69 samples were found to be strongly associated with either the high or low potency responses. Five sites were in the CD4-induced coreceptor binding site of gp120, suggesting an important role for this region in the elicitation of broadly neutralizing antibody responses against HIV-1.

Task-driven adaptive statistical compressive sensing of gaussian mixture models

A framework for adaptive and non-adaptive statistical compressive sensing is developed, where a statistical model replaces the standard sparsity model of classical compressive sensing. We propose within this framework optimal task-specific sensing protocols specifically and jointly designed for classification and reconstruction. A two-step adaptive sensing paradigm is developed, where online sensing is applied to detect the signal class in the first step, followed by a reconstruction step adapted to the detected class and the observed samples. The approach is based on information theory, here tailored for Gaussian mixture models (GMMs), where an information-theoretic objective relationship between the sensed signals and a representation of the specific task of interest is maximized. Experimental results using synthetic signals, Landsat satellite attributes, and natural images of different sizes and with different noise levels show the improvements achieved using the proposed framework when compared to more standard sensing protocols. The underlying formulation can be applied beyond GMMs, at the price of higher mathematical and computational complexity. © 1991-2012 IEEE.

Decoding an olfactory mechanism of kin recognition and inbreeding avoidance in a primate.

BACKGROUND: Like other vertebrates, primates recognize their relatives, primarily to minimize inbreeding, but also to facilitate nepotism. Although associative, social learning is typically credited for discrimination of familiar kin, discrimination of unfamiliar kin remains unexplained. As sex-biased dispersal in long-lived species cannot consistently prevent encounters between unfamiliar kin, inbreeding remains a threat and mechanisms to avoid it beg explanation. Using a molecular approach that combined analyses of biochemical and microsatellite markers in 17 female and 19 male ring-tailed lemurs (Lemur catta), we describe odor-gene covariance to establish the feasibility of olfactory-mediated kin recognition. RESULTS: Despite derivation from different genital glands, labial and scrotal secretions shared about 170 of their respective 338 and 203 semiochemicals. In addition, these semiochemicals encoded information about genetic relatedness within and between the sexes. Although the sexes showed opposite seasonal patterns in signal complexity, the odor profiles of related individuals (whether same-sex or mixed-sex dyads) converged most strongly in the competitive breeding season. Thus, a strong, mutual olfactory signal of genetic relatedness appeared specifically when such information would be crucial for preventing inbreeding. That weaker signals of genetic relatedness might exist year round could provide a mechanism to explain nepotism between unfamiliar kin. CONCLUSION: We suggest that signal convergence between the sexes may reflect strong selective pressures on kin recognition, whereas signal convergence within the sexes may arise as its by-product or function independently to prevent competition between unfamiliar relatives. The link between an individual's genome and its olfactory signals could be mediated by biosynthetic pathways producing polymorphic semiochemicals or by carrier proteins modifying the individual bouquet of olfactory cues. In conclusion, we unveil a possible olfactory mechanism of kin recognition that has specific relevance to understanding inbreeding avoidance and nepotistic behavior observed in free-ranging primates, and broader relevance to understanding the mechanisms of vertebrate olfactory communication.

Interpreting Human Genetic Variation through Genomics and In Vivo Models

Improvements in genomic technology, both in the increased speed and reduced cost of sequencing, have expanded the appreciation of the abundance of human genetic variation. However the sheer amount of variation, as well as the varying type and genomic content of variation, poses a challenge in understanding the clinical consequence of a single mutation. This work uses several methodologies to interpret the observed variation in the human genome, and presents novel strategies for the prediction of allele pathogenicity.

Using the zebrafish model system as an in vivo assay of allele function, we identified a novel driver of Bardet-Biedl Syndrome (BBS) in CEP76. A combination of targeted sequencing of 785 cilia-associated genes in a cohort of BBS patients and subsequent in vivo functional assays recapitulating the human phenotype gave strong evidence for the role of CEP76 mutations in the pathology of an affected family. This portion of the work demonstrated the necessity of functional testing in validating disease-associated mutations, and added to the catalogue of known BBS disease genes.

Further study into the role of copy-number variations (CNVs) in a cohort of BBS patients showed the significant contribution of CNVs to disease pathology. Using high-density array comparative genomic hybridization (aCGH) we were able to identify pathogenic CNVs as small as several hundred bp. Dissection of constituent gene and in vivo experiments investigating epistatic interactions between affected genes allowed for an appreciation of several paradigms by which CNVs can contribute to disease. This study revealed that the contribution of CNVs to disease in BBS patients is much higher than previously expected, and demonstrated the necessity of consideration of CNV contribution in future (and retrospective) investigations of human genetic disease.

Finally, we used a combination of comparative genomics and in vivo complementation assays to identify second-site compensatory modification of pathogenic alleles. These pathogenic alleles, which are found compensated in other species (termed compensated pathogenic deviations [CPDs]), represent a significant fraction (from 3 – 10%) of human disease-associated alleles. In silico pathogenicity prediction algorithms, a valuable method of allele prioritization, often misrepresent these alleles as benign, leading to omission of possibly informative variants in studies of human genetic disease. We created a mathematical model that was able to predict CPDs and putative compensatory sites, and functionally showed in vivo that second-site mutation can mitigate the pathogenicity of disease alleles. Additionally, we made publically available an in silico module for the prediction of CPDs and modifier sites.

These studies have advanced the ability to interpret the pathogenicity of multiple types of human variation, as well as made available tools for others to do so as well.

Coding Strategies and Implementations of Compressive Sensing

This dissertation studies the coding strategies of computational imaging to overcome the limitation of conventional sensing techniques. The information capacity of conventional sensing is limited by the physical properties of optics, such as aperture size, detector pixels, quantum efficiency, and sampling rate. These parameters determine the spatial, depth, spectral, temporal, and polarization sensitivity of each imager. To increase sensitivity in any dimension can significantly compromise the others.

This research implements various coding strategies subject to optical multidimensional imaging and acoustic sensing in order to extend their sensing abilities. The proposed coding strategies combine hardware modification and signal processing to exploiting bandwidth and sensitivity from conventional sensors. We discuss the hardware architecture, compression strategies, sensing process modeling, and reconstruction algorithm of each sensing system.

Optical multidimensional imaging measures three or more dimensional information of the optical signal. Traditional multidimensional imagers acquire extra dimensional information at the cost of degrading temporal or spatial resolution. Compressive multidimensional imaging multiplexes the transverse spatial, spectral, temporal, and polarization information on a two-dimensional (2D) detector. The corresponding spectral, temporal and polarization coding strategies adapt optics, electronic devices, and designed modulation techniques for multiplex measurement. This computational imaging technique provides multispectral, temporal super-resolution, and polarization imaging abilities with minimal loss in spatial resolution and noise level while maintaining or gaining higher temporal resolution. The experimental results prove that the appropriate coding strategies may improve hundreds times more sensing capacity.

Human auditory system has the astonishing ability in localizing, tracking, and filtering the selected sound sources or information from a noisy environment. Using engineering efforts to accomplish the same task usually requires multiple detectors, advanced computational algorithms, or artificial intelligence systems. Compressive acoustic sensing incorporates acoustic metamaterials in compressive sensing theory to emulate the abilities of sound localization and selective attention. This research investigates and optimizes the sensing capacity and the spatial sensitivity of the acoustic sensor. The well-modeled acoustic sensor allows localizing multiple speakers in both stationary and dynamic auditory scene; and distinguishing mixed conversations from independent sources with high audio recognition rate.