Cross-dataset Analysis Research Articles

Indigeneity in Context—Evolving Maya Ch’orti’ Notions of Cultural Identity: A Qualitative Study From 1993 to 2019

The Maya Ch’orti’ people are Indigenous to the border regions of Honduras and Guatemala. For a great number of years, they have faced structural racism and discrimination combined with continued land dispossession. In the contexts of colonization and the creation of the modern Honduran state, any identification with Indigenous groups has generally had negative connotations. Since the advent of the Indigenous Peoples’ Rights Discourses (IPRD), however, more communities have actively aligned with the Maya Ch’orti’ people. In our investigation of the ways in which cultural identity and belonging to a Maya Ch’orti’ community have shifted and evolved over time, we analyzed a total of 45 interview transcripts across three time points (1990s: n = 15, 2003–2004: n = 15, and 2018–2019: n = 15) using thematic analysis and a cross–data set analysis. Five major themes were developed: “Our land,” “In the past, we were oppressed and ashamed,” “Difficult present life: We are still oppressed,” “Maya Ch’orti’ revitalization and empowerment,” and “Being Indigenous.” As we compared these themes across our time points, it became apparent that notions of cultural identity have evolved in the Maya Ch’orti’ community from a negative appraisal to a positive, empowered, and confident notion. Our findings are being discussed in connection to the context of the IPRD and Indigenous peoples in other countries.

Evaluation of Normalization Algorithms for Breast Mammogram Mass Segmentation

Medical images acquired at different institutions have different data distributions due to varying scanners, imaging protocols, or patient cohorts. Thus, normalizing samples before using them to train the Deep Learning (DL) models is an essential task. Scholars have used various normalization methods to normalize mammograms. However, which normalization method works best on mammograms and if it can help training generalize the DL model has not been analyzed in the literature. This paper presents a comparative analysis of commonly used normalization methods, such as scaling, Contrast Limited Adaptive Histogram Equalization (CLAHE), and Z-score normalization on mammograms for breast mass segmentation tasks. We trained the Attention Unet model on INbreast and the subset of the Cohort of Screen-Aged Women dataset (CSAW-S). We performed cross-dataset analysis on trained models to investigate how normalization methods helped train robust models to handle diverse datasets. Our results suggest that the normalizing method of input data affects the segmentation model’s performance. We achieved the best scores with the Z-score normalization closely followed by CLAHE. However, further investigation suggests that the normalization method’s performance may depend on the variability of mammograms in the dataset. Results of cross-dataset analysis suggest the importance of choosing the normalization method for consistent DL model performance. To our knowledge, this study represents the first systematic analysis of the performance of prevalent normalization techniques on a diverse dataset of mammogram images. The insights gained from our experiments can guide the selection of normalization methods for improved mass segmentation accuracy.

Markov-based Neural Networks for Heart Sound Segmentation: Using domain knowledge in a principled way.

This work considers the problem of segmenting heart sounds into their fundamental components. We unify statistical and data-driven solutions by introducing Markov-based Neural Networks (MNNs), a hybrid end-to-end framework that exploits Markov models as statistical inductive biases for an Artificial Neural Network (ANN) discriminator. We show that an MNN leveraging a simple one-dimensional Convolutional ANN significantly outperforms two recent purely data-driven solutions for this task in two publicly available datasets: PhysioNet 2016 (Sensitivity: 0.947 ±0.02; Positive Predictive Value : 0.937 ±0.025) and the CirCor DigiScope 2022 (Sensitivity: 0.950 ±0.008; Positive Predictive Value: 0.943 ±0.012). We also propose a novel gradient-based unsupervised learning algorithm that effectively makes the MNN adaptive to unseen datum sampled from unknown distributions. We perform a cross dataset analysis and show that an MNN pre-trained in the CirCor DigiScope 2022 can benefit from an average improvement of 3.90% Positive Predictive Value on unseen observations from the PhysioNet 2016 dataset using this method.

Towards Equitable Healthcare: A Cross Dataset Analysis of Healthcare and Telehealth Access

ABSTRACTHistory has repeatedly shown that when it comes to healthcare, there exist significant disparities across different sub‐populations. Part of this problem is challenges with access to healthcare services and providers. In recent years, we have seen solutions emerge that attempt to address this issue, namely telehealth. But little to no work has been done to glean insights into factors contributing to lower access and whether telehealth adequately supports populations that struggle with physical access issues. In this paper, we present findings from an exploratory, cross‐dataset analysis of healthcare access in Virginia, USA. We conducted quantitative and qualitative analyses lyses to determine correlations between socioeconomic factors and physical access. We found that population density significantly impacts physical access to healthcare. We also found that lower income communities in rural areas appear to be less satisfied with the quality of care and less likely to have adequate access to alternative options such as telehealth services. Our work has implications for research aimed at understanding healthcare access and practice that aims to derive solutions to close the healthcare access gap.

Abstract 5380: Systematic evaluation and comparison of drug response prediction models: a case study of prediction generalization across cell lines datasets

Abstract Predictive modeling holds great promise for improving personalized cancer treatment and efficiency of drug development. In recent years, deep learning (DL) has been extensively explored for drug response prediction (DRP), outperforming classical machine learning in prediction generalization to new data. Despite the considerable interest in DRP, no agreed-upon methodology for evaluating and comparing the diverse DL models yet exists. Existing papers generally demonstrate the performance of proposed models using cross-validation within a single cell line dataset and compare with baseline models of their choice, substantially limiting the scope and validity of model evaluation and comparison. In this work, we investigate the ability of DRP models for generalizing predictions across datasets of multiple drug screening studies, a more challenging scenario mimicking practical applications of DRP models. Five cell line datasets and six community DRP models with advanced DL architectures have been explored. Public cell line drug screening datasets have been curated and processed for this analysis, including CCLE, CTRP, GDSC1, GDSC2, and GCSI. For each dataset, the same preprocessing pipeline was used to generate cell line gene expressions, drug representations, and drug response values. The six DRP models include advanced architectures and feature engineering methods such as transformer, graph neural network, and image representation of tabular data. Systematic model curation and training have been applied, including consistent training and testing data splits across models and hyperparameter optimization (HPO). To cope with the large-scale model training and HPO, automatic workflows have been implemented and executed on high-performance computing systems. A 5-by-5 matrix of prediction scores, corresponding to the five datasets in both row and column dimensions, has been generated for each model, with off-diagonal values representing the cross-dataset generalization. Despite the advanced DL techniques, all models exhibit substantially inferior performance in cross-dataset analysis as compared with cross-validation within a single dataset. This result demonstrates the challenge of cross-dataset generalization for DRP and motivates the need for rigorous and systematic evaluation of DRP models, which simulates real-world applications. Citation Format: Alexander Partin, Thomas S. Brettin, Yitan Zhu, Jamie Overbeek, Oleksandr Narykov, Priyanka Vasanthakumari, Austin Clyde, Sara E. Jones, Satishkumar Ranganathan Ganakammal, Justin M. Wozniak, Andreas Wilke, Jamaludin Mohd-Yusof, Michael R. Weil, Alexander T. Pearson, Rick L. Stevens. Systematic evaluation and comparison of drug response prediction models: a case study of prediction generalization across cell lines datasets. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5380.

Cross Dataset Analysis for Generalizability of HRV-Based Stress Detection Models.

Stress is an increasingly prevalent mental health condition across the world. In Europe, for example, stress is considered one of the most common health problems, and over USD 300 billion are spent on stress treatments annually. Therefore, monitoring, identification and prevention of stress are of the utmost importance. While most stress monitoring is carried out through self-reporting, there are now several studies on stress detection from physiological signals using Artificial Intelligence algorithms. However, the generalizability of these models is only rarely discussed. The main goal of this work is to provide a monitoring proof-of-concept tool exploring the generalization capabilities of Heart Rate Variability-based machine learning models. To this end, two Machine Learning models are used, Logistic Regression and Random Forest to analyze and classify stress in two datasets differing in terms of protocol, stressors and recording devices. First, the models are evaluated using leave-one-subject-out cross-validation with train and test samples from the same dataset. Next, a cross-dataset validation of the models is performed, that is, leave-one-subject-out models trained on a Multi-modal Dataset for Real-time, Continuous Stress Detection from Physiological Signals dataset and validated using the University of Waterloo stress dataset. While both logistic regression and random forest models achieve good classification results in the independent dataset analysis, the random forest model demonstrates better generalization capabilities with a stable F1 score of 61%. This indicates that the random forest can be used to generalize HRV-based stress detection models, which can lead to better analyses in the mental health and medical research field through training and integrating different models.

The value of cross-data set analysis for automobile insurance fraud detection

This study focuses on personal automobile policies underwritten. Its aim is to provide decision support and to apply new models with good predictive performance and high operational efficiency. We propose a new approach by constructing a score that evolves over the life of a claim. It consists of creating a score at the opening of a claim and another derived from the information of the first adjuster’s report. Natural language processing is also used on a textual variable relating to the description of the claim provided by the agency. The fraud score is estimated by using a gradient boosting machine (GBM) and a neural network. The results are interpreted using the local interpretable model-agnostic explanations (LIME). They show that fraud detection is improved when all the information and the textual variable are included. Furthermore, we observe that the GBM method overperforms the neural network approach.

ViolenceNet: Dense Multi-Head Self-Attention with Bidirectional Convolutional LSTM for Detecting Violence

Introducing efficient automatic violence detection in video surveillance or audiovisual content monitoring systems would greatly facilitate the work of closed-circuit television (CCTV) operators, rating agencies or those in charge of monitoring social network content. In this paper we present a new deep learning architecture, using an adapted version of DenseNet for three dimensions, a multi-head self-attention layer and a bidirectional convolutional long short-term memory (LSTM) module, that allows encoding relevant spatio-temporal features, to determine whether a video is violent or not. Furthermore, an ablation study of the input frames, comparing dense optical flow and adjacent frames subtraction and the influence of the attention layer is carried out, showing that the combination of optical flow and the attention mechanism improves results up to 4.4%. The conducted experiments using four of the most widely used datasets for this problem, matching or exceeding in some cases the results of the state of the art, reducing the number of network parameters needed (4.5 millions), and increasing its efficiency in test accuracy (from 95.6% on the most complex dataset to 100% on the simplest one) and inference time (less than 0.3 s for the longest clips). Finally, to check if the generated model is able to generalize violence, a cross-dataset analysis is performed, which shows the complexity of this approach: using three datasets to train and testing on the remaining one the accuracy drops in the worst case to 70.08% and in the best case to 81.51%, which points to future work oriented towards anomaly detection in new datasets.

Resting-state functional connectivity of the human hippocampus in periadolescent children: Associations with age and memory performance.

The hippocampus is necessary for declarative (relational) memory, and the ability to form hippocampal‐dependent memories develops through late adolescence. This developmental trajectory of hippocampal‐dependent memory could reflect maturation of intrinsic functional brain networks, but resting‐state functional connectivity (rs‐FC) of the human hippocampus is not well‐characterized for periadolescent children. Measuring hippocampal rs‐FC in periadolescence would thus fill a gap, and testing covariance of hippocampal rs‐FC with age and memory could inform theories of cognitive development. Here, we studied hippocampal rs‐FC in a cross‐sectional sample of healthy children (N = 96; 59 F; age 9–15 years) using a seed‐based approach, and linked these data with NIH Toolbox measures, the Picture‐Sequence Memory Test (PSMT) and the List Sorting Working Memory Test (LSWMT). The PSMT was expected to rely more on hippocampal‐dependent memory than the LSWMT. We observed hippocampal rs‐FC with an extensive brain network including temporal, parietal, and frontal regions. This pattern was consistent with prior work measuring hippocampal rs‐FC in younger and older samples. We also observed novel, regionally specific variation in hippocampal rs‐FC with age and hippocampal‐dependent memory but not working memory. Evidence consistent with these findings was observed in a second, validation dataset of similar‐age healthy children drawn from the Philadelphia Neurodevelopment Cohort. Further, a cross‐dataset analysis suggested generalizable properties of hippocampal rs‐FC and covariance with age and memory. Our findings connect prior work by describing hippocampal rs‐FC and covariance with age and memory in typically developing periadolescent children, and our observations suggest a developmental trajectory for brain networks that support hippocampal‐dependent memory.

The FaceChannel: A Fast and Furious Deep Neural Network for Facial Expression Recognition

Current state-of-the-art models for automatic facial expression recognition (FER) are based on very deep neural networks that are effective but rather expensive to train. Given the dynamic conditions of FER, this characteristic hinders such models of been used as a general affect recognition. In this paper, we address this problem by formalizing the FaceChannel, a light-weight neural network that has much fewer parameters than common deep neural networks. We introduce an inhibitory layer that helps to shape the learning of facial features in the last layer of the network and, thus, improving performance while reducing the number of trainable parameters. To evaluate our model, we perform a series of experiments on different benchmark datasets and demonstrate how the FaceChannel achieves a comparable, if not better, performance to the current state-of-the-art in FER. Our experiments include cross-dataset analysis, to estimate how our model behaves on different affective recognition conditions. We conclude our paper with an analysis of how FaceChannel learns and adapts the learned facial features towards the different datasets.

Exploring Integrative Analysis Using the BioMedical Evidence Graph.

PURPOSEThe analysis of cancer biology data involves extremely heterogeneous data sets, including information from RNA sequencing, genome-wide copy number, DNA methylation data reporting on epigenetic regulation, somatic mutations from whole-exome or whole-genome analyses, pathology estimates from imaging sections or subtyping, drug response or other treatment outcomes, and various other clinical and phenotypic measurements. Bringing these different resources into a common framework, with a data model that allows for complex relationships as well as dense vectors of features, will unlock integrated data set analysis.METHODSWe introduce the BioMedical Evidence Graph (BMEG), a graph database and query engine for discovery and analysis of cancer biology. The BMEG is unique from other biologic data graphs in that sample-level molecular and clinical information is connected to reference knowledge bases. It combines gene expression and mutation data with drug-response experiments, pathway information databases, and literature-derived associations.RESULTSThe construction of the BMEG has resulted in a graph containing > 41 million vertices and 57 million edges. The BMEG system provides a graph query–based application programming interface to enable analysis, with client code available for Python, Javascript, and R, and a server online at bmeg.io. Using this system, we have demonstrated several forms of cross–data set analysis to show the utility of the system.CONCLUSIONThe BMEG is an evolving resource dedicated to enabling integrative analysis. We have demonstrated queries on the system that illustrate mutation significance analysis, drug-response machine learning, patient-level knowledge-base queries, and pathway level analysis. We have compared the resulting graph to other available integrated graph systems and demonstrated the former is unique in the scale of the graph and the type of data it makes available.

COVID-19 detection in CT images with deep learning: A voting-based scheme and cross-datasets analysis.

Early detection and diagnosis are critical factors to control the COVID-19 spreading. A number of deep learning-based methodologies have been recently proposed for COVID-19 screening in CT scans as a tool to automate and help with the diagnosis. These approaches, however, suffer from at least one of the following problems: (i) they treat each CT scan slice independently and (ii) the methods are trained and tested with sets of images from the same dataset. Treating the slices independently means that the same patient may appear in the training and test sets at the same time which may produce misleading results. It also raises the question of whether the scans from the same patient should be evaluated as a group or not. Moreover, using a single dataset raises concerns about the generalization of the methods. Different datasets tend to present images of varying quality which may come from different types of CT machines reflecting the conditions of the countries and cities from where they come from. In order to address these two problems, in this work, we propose an Efficient Deep Learning Technique for the screening of COVID-19 with a voting-based approach. In this approach, the images from a given patient are classified as group in a voting system. The approach is tested in the two biggest datasets of COVID-19 CT analysis with a patient-based split. A cross dataset study is also presented to assess the robustness of the models in a more realistic scenario in which data comes from different distributions. The cross-dataset analysis has shown that the generalization power of deep learning models is far from acceptable for the task since accuracy drops from 87.68% to 56.16% on the best evaluation scenario. These results highlighted that the methods that aim at COVID-19 detection in CT-images have to improve significantly to be considered as a clinical option and larger and more diverse datasets are needed to evaluate the methods in a realistic scenario.

Toward Unbiased Facial Expression Recognition in the Wild via Cross-Dataset Adaptation

Despite various success in computer vision with facial images (e.g., face detection, recognition, and generation), facial expression recognition is still a challenging problem yet to be solved. This is because of simple but fundamental bottlenecks: 1) no global agreement on different facial expressions, 2) significant dataset biases that prevent cross-dataset analysis for a large-scale study, and 3) high class imbalance in in-the-wild datasets that causes inconsistency in predicting expressions in images using a machine learning algorithm. To tackle these issues, we propose a novel Deep Learning approach via adaptive cross-dataset scheme. We combine multiple in-the-wild datasets to secure sufficient training samples while minimizing dataset bias using ideas of reversal gradients to retain generality. For this, we introduce a flexible objective function that can control for skewed label distributions in the dataset. Incorporating these ideas, together with the ResNet pipeline as a backbone, we carried extensive experiments to validate our ideas using three independent in-the-wild facial expression datasets, which first confirmed bias from different datasets and yielded improved performance on facial expression recognition using the multi-site dataset.

A multitask clustering approach for single-cell RNA-seq analysis in Recessive Dystrophic Epidermolysis Bullosa.

Single-cell RNA sequencing (scRNA-seq) has been widely applied to discover new cell types by detecting sub-populations in a heterogeneous group of cells. Since scRNA-seq experiments have lower read coverage/tag counts and introduce more technical biases compared to bulk RNA-seq experiments, the limited number of sampled cells combined with the experimental biases and other dataset specific variations presents a challenge to cross-dataset analysis and discovery of relevant biological variations across multiple cell populations. In this paper, we introduce a method of variance-driven multitask clustering of single-cell RNA-seq data (scVDMC) that utilizes multiple single-cell populations from biological replicates or different samples. scVDMC clusters single cells in multiple scRNA-seq experiments of similar cell types and markers but varying expression patterns such that the scRNA-seq data are better integrated than typical pooled analyses which only increase the sample size. By controlling the variance among the cell clusters within each dataset and across all the datasets, scVDMC detects cell sub-populations in each individual experiment with shared cell-type markers but varying cluster centers among all the experiments. Applied to two real scRNA-seq datasets with several replicates and one large-scale droplet-based dataset on three patient samples, scVDMC more accurately detected cell populations and known cell markers than pooled clustering and other recently proposed scRNA-seq clustering methods. In the case study applied to in-house Recessive Dystrophic Epidermolysis Bullosa (RDEB) scRNA-seq data, scVDMC revealed several new cell types and unknown markers validated by flow cytometry. MATLAB/Octave code available at https://github.com/kuanglab/scVDMC.

An integrative analysis of regional gene expression profiles in the human brain

Studies of the brain’s transcriptome have become prominent in recent years, resulting in an accumulation of datasets with somewhat distinct attributes. These datasets, which are often analyzed only in isolation, also are often collected with divergent goals, which are reflected in their sampling properties. While many researchers have been interested in sampling gene expression in one or a few brain areas in a large number of subjects, recent efforts from the Allen Institute for Brain Sciences and others have focused instead on dense neuroanatomical sampling, necessarily limiting the number of individual donor brains studied. The purpose of the present work is to develop methods that draw on the complementary strengths of these two types of datasets for study of the human brain, and to characterize the anatomical specificity of gene expression profiles and gene co-expression networks derived from human brains using different specific technologies. The approach is applied using two publicly accessible datasets: (1) the high anatomical resolution Allen Human Brain Atlas (AHBA, Hawrylycz et al., 2012) and (2) a relatively large sample size, but comparatively coarse neuroanatomical dataset described previously by Gibbs et al. (2010). We found a relatively high degree of correspondence in differentially expressed genes and regional gene expression profiles across the two datasets. Gene co-expression networks defined in individual brain regions were less congruent, but also showed modest anatomical specificity. Using gene modules derived from the Gibbs dataset and from curated gene lists, we demonstrated varying degrees of anatomical specificity based on two classes of methods, one focused on network modularity and the other focused on enrichment of expression levels. Two approaches to assessing the statistical significance of a gene set’s modularity in a given brain region were studied, which provide complementary information about the anatomical specificity of a gene network of interest. Overall, the present work demonstrates the feasibility of cross-dataset analysis of human brain microarray studies, and offers a new approach to annotating gene lists in a neuroanatomical context.

The CARMEN data sharing portal project: what have we learned?

The CARMEN data sharing portal project: what have we learned?

Points of Interest and Visual Dictionaries for Automatic Retinal Lesion Detection

In this paper, we present an algorithm to detect the presence of diabetic retinopathy (DR)-related lesions from fundus images based on a common analytical approach that is capable of identifying both red and bright lesions without requiring specific pre- or postprocessing. Our solution constructs a visual word dictionary representing points of interest (PoIs) located within regions marked by specialists that contain lesions associated with DR and classifies the fundus images based on the presence or absence of these PoIs as normal or DR-related pathology. The novelty of our approach is in locating DR lesions in the optic fundus images using visual words that combines feature information contained within the images in a framework easily extendible to different types of retinal lesions or pathologies and builds a specific projection space for each class of interest (e.g., white lesions such as exudates or normal regions) instead of a common dictionary for all classes. The visual words dictionary was applied to classifying bright and red lesions with classical cross validation and cross dataset validation to indicate the robustness of this approach. We obtained an area under the curve (AUC) of 95.3% for white lesion detection and an AUC of 93.3% for red lesion detection using fivefold cross validation and our own data consisting of 687 images of normal retinae, 245 images with bright lesions, 191 with red lesions, and 109 with signs of both bright and red lesions. For cross dataset analysis, the visual dictionary also achieves compelling results using our images as the training set and the RetiDB and Messidor images as test sets. In this case, the image classification resulted in an AUC of 88.1% when classifying the RetiDB dataset and in an AUC of 89.3% when classifying the Messidor dataset, both cases for bright lesion detection. The results indicate the potential for training with different acquisition images under different setup conditions with a high accuracy of referral based on the presence of either red or bright lesions or both. The robustness of the visual dictionary against image quality (blurring), resolution, and retinal background, makes it a strong candidate for DR screening of large, diverse communities with varying cameras and settings and levels of expertise for image capture.

DiseaseMeth: a human disease methylation database

DNA methylation is an important epigenetic modification for genomic regulation in higher organisms that plays a crucial role in the initiation and progression of diseases. The integration and mining of DNA methylation data by methylation-specific PCR and genome-wide profiling technology could greatly assist the discovery of novel candidate disease biomarkers. However, this is difficult without a comprehensive DNA methylation repository of human diseases. Therefore, we have developed DiseaseMeth, a human disease methylation database (http://bioinfo.hrbmu.edu.cn/diseasemeth). Its focus is the efficient storage and statistical analysis of DNA methylation data sets from various diseases. Experimental information from over 14 000 entries and 175 high-throughput data sets from a wide number of sources have been collected and incorporated into DiseaseMeth. The latest release incorporates the gene-centric methylation data of 72 human diseases from a variety of technologies and platforms. To facilitate data extraction, DiseaseMeth supports multiple search options such as gene ID and disease name. DiseaseMeth provides integrated gene methylation data based on cross-data set analysis for disease and normal samples. These can be used for in-depth identification of differentially methylated genes and the investigation of gene–disease relationship.

Identification of robust, clinically relevant phenotypes of breast cancer from genome scale molecular profiling.

Abstract Abstract #2036 Genome scale analysis by microarrays or array CGH holds the promise to identify novel subclasses of breast cancer or characterize functional modules of clinical relevance in a robust fashion. However, the success of these endeavors is highly dependent on several factors such as the general noise structure and noise level of high throughput measurements or the strength of association between biomarkers/gene modules and clinical outcome. We will address two issues deeply rooted in the high throughput nature of genome scale profiling, which are relevant for the accurate analysis of clinical microarray data sets: systematic bias in clinical microarray data and establishing strategies that allow extracting robust, convergent information/classification from genome scale molecular profiling of breast cancer.
 As will be demonstrated, clinical microarray data are burdened with a high level of systematic bias. We identified sources of technical bias affecting many genes in concert, thus causing spurious correlations in clinical data sets and false associations between genes and clinical variables. A method will be presented to correct for technical bias in clinical microarray data, which increased concordance with known biological relationships in multiple data sets.
 Gene expression profiling based classification of breast cancer and prognostic or clinical response associated gene expression signatures are usually derived from a single data set. However, any result extracted from a single data set will reflect to a large extent the technical (which genes are measured reliably on the microarray) and biological (such as cohort selection) bias of the given cohort. An alternative approach uses multiple, in this case 5 different, analogous clinical data sets and determines the robust, convergent information emerging from the cross data-set analysis. We will present such a method, which will reduce the impact of data set specific bias and outline robust functional modules in breast cancer, ultimately leading to the reevaluation of gene expression profiling based subtyping and diagnostic gene expression signatures in breast cancer.
 Finally, we will present evidence for the existence of at least two fundamentally different types of genome instability in breast cancer with direct implications for response to chemotherapy. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 2036.