A generative vision model that trains with high data efficiency and breaks text-based CAPTCHAs.

Learning from a few examples and generalizing to markedly different situations are capabilities of human visual intelligence that are yet to be matched by leading machine learning models. By drawing inspiration from systems neuroscience, we introduce a probabilistic generative model for vision in which message-passing-based inference handles recognition, segmentation, and reasoning in a unified way. The model demonstrates excellent generalization and occlusion-reasoning capabilities and outperforms deep neural networks on a challenging scene text recognition benchmark while being 300-fold more data efficient. In addition, the model fundamentally breaks the defense of modern text-based CAPTCHAs (Completely Automated Public Turing test to tell Computers and Humans Apart) by generatively segmenting characters without CAPTCHA-specific heuristics. Our model emphasizes aspects such as data efficiency and compositionality that may be important in the path toward general artificial intelligence.

Quantitative Analysis of Protein Expression to Study Lineage Specification in Mouse Preimplantation Embryos.

This protocol presents a method to perform quantitative, single-cell in situ analyses of protein expression to study lineage specification in mouse preimplantation embryos. The procedures necessary for embryo collection, immunofluorescence, imaging on a confocal microscope, and image segmentation and analysis are described. This method allows quantitation of the expression of multiple nuclear markers and the spatial (XYZ) coordinates of all cells in the embryo. It takes advantage of MINS, an image segmentation software tool specifically developed for the analysis of confocal images of preimplantation embryos and embryonic stem cell (ESC) colonies. MINS carries out unsupervised nuclear segmentation across the X, Y and Z dimensions, and produces information on cell position in three-dimensional space, as well as nuclear fluorescence levels for all channels with minimal user input. While this protocol has been optimized for the analysis of images of preimplantation stage mouse embryos, it can easily be adapted to the analysis of any other samples exhibiting a good signal-to-noise ratio and where high nuclear density poses a hurdle to image segmentation (e.g., expression analysis of embryonic stem cell (ESC) colonies, differentiating cells in culture, embryos of other species or stages, etc.).

Active structured learning for cell tracking: algorithm, framework, and usability.

One distinguishing property of life is its temporal dynamics, and it is hence only natural that time lapse experiments play a crucial role in modern biomedical research areas such as signaling pathways, drug discovery or developmental biology. Such experiments yield a very large number of images that encode complex cellular activities, and reliable automated cell tracking emerges naturally as a prerequisite for further quantitative analysis. However, many existing cell tracking methods are restricted to using only a small number of features to allow for manual tweaking. In this paper, we propose a novel cell tracking approach that embraces a powerful machine learning technique to optimize the tracking parameters based on user annotated tracks. Our approach replaces the tedious parameter tuning with parameter learning and allows for the use of a much richer set of complex tracking features, which in turn affords superior prediction accuracy. Furthermore, we developed an active learning approach for efficient training data retrieval, which reduces the annotation effort to only 17%. In practical terms, our approach allows life science researchers to inject their expertise in a more intuitive and direct manner. This process is further facilitated by using a glyph visualization technique for ground truth annotation and validation. Evaluation and comparison on several publicly available benchmark sequences show significant performance improvement over recently reported approaches. Code and software tools are provided to the public.

A rapid and efficient 2D/3D nuclear segmentation method for analysis of early mouse embryo and stem cell image data.

Segmentation is a fundamental problem that dominates the success of microscopic image analysis. In almost 25 years of cell detection software development, there is still no single piece of commercial software that works well in practice when applied to early mouse embryo or stem cell image data. To address this need, we developed MINS (modular interactive nuclear segmentation) as a MATLAB/C++-based segmentation tool tailored for counting cells and fluorescent intensity measurements of 2D and 3D image data. Our aim was to develop a tool that is accurate and efficient yet straightforward and user friendly. The MINS pipeline comprises three major cascaded modules: detection, segmentation, and cell position classification. An extensive evaluation of MINS on both 2D and 3D images, and comparison to related tools, reveals improvements in segmentation accuracy and usability. Thus, its accuracy and ease of use will allow MINS to be implemented for routine single-cell-level image analyses.

Hexicon 2: automated processing of hydrogen-deuterium exchange mass spectrometry data with improved deuteration distribution estimation.

Hydrogen-deuterium exchange (HDX) experiments analyzed by mass spectrometry (MS) provide information about the dynamics and the solvent accessibility of protein backbone amide hydrogen atoms. Continuous improvement of MS instrumentation has contributed to the increasing popularity of this method; however, comprehensive automated data analysis is only beginning to mature. We present Hexicon 2, an automated pipeline for data analysis and visualization based on the previously published program Hexicon (Lou et al. 2010). Hexicon 2 employs the sensitive NITPICK peak detection algorithm of its predecessor in a divide-and-conquer strategy and adds new features, such as chromatogram alignment and improved peptide sequence assignment. The unique feature of deuteration distribution estimation was retained in Hexicon 2 and improved using an iterative deconvolution algorithm that is robust even to noisy data. In addition, Hexicon 2 provides a data browser that facilitates quality control and provides convenient access to common data visualization tasks. Analysis of a benchmark dataset demonstrates superior performance of Hexicon 2 compared with its predecessor in terms of deuteration centroid recovery and deuteration distribution estimation. Hexicon 2 greatly reduces data analysis time compared with manual analysis, whereas the increased number of peptides provides redundant coverage of the entire protein sequence. Hexicon 2 is a standalone application available free of charge under http://hx2.mpimf-heidelberg.mpg.de.

Oct4 is required for lineage priming in the developing inner cell mass of the mouse blastocyst.

The transcription factor Oct4 is required in vitro for establishment and maintenance of embryonic stem cells and for reprogramming somatic cells to pluripotency. In vivo, it prevents the ectopic differentiation of early embryos into trophoblast. Here, we further explore the role of Oct4 in blastocyst formation and specification of epiblast versus primitive endoderm lineages using conditional genetic deletion. Experiments involving mouse embryos deficient for both maternal and zygotic Oct4 suggest that it is dispensable for zygote formation, early cleavage and activation of Nanog expression. Nanog protein is significantly elevated in the presumptive inner cell mass of Oct4 null embryos, suggesting an unexpected role for Oct4 in attenuating the level of Nanog, which might be significant for priming differentiation during epiblast maturation. Induced deletion of Oct4 during the morula to blastocyst transition disrupts the ability of inner cell mass cells to adopt lineage-specific identity and acquire the molecular profile characteristic of either epiblast or primitive endoderm. Sox17, a marker of primitive endoderm, is not detected following prolonged culture of such embryos, but can be rescued by provision of exogenous FGF4. Interestingly, functional primitive endoderm can be rescued in Oct4-deficient embryos in embryonic stem cell complementation assays, but only if the host embryos are at the pre-blastocyst stage. We conclude that cell fate decisions within the inner cell mass are dependent upon Oct4 and that Oct4 is not cell-autonomously required for the differentiation of primitive endoderm derivatives, as long as an appropriate developmental environment is established.

Graph-regularized 3D shape reconstruction from highly anisotropic and noisy images.

Analysis of microscopy images can provide insight into many biological processes. One particularly challenging problem is cellular nuclear segmentation in highly anisotropic and noisy 3D image data. Manually localizing and segmenting each and every cellular nucleus is very time-consuming, which remains a bottleneck in large-scale biological experiments. In this work, we present a tool for automated segmentation of cellular nuclei from 3D fluorescent microscopic data. Our tool is based on state-of-the-art image processing and machine learning techniques and provides a user-friendly graphical user interface. We show that our tool is as accurate as manual annotation and greatly reduces the time for the registration.

Live imaging, identifying, and tracking single cells in complex populations in vivo and ex vivo.

Advances in optical imaging technologies combined with the use of genetically encoded fluorescent proteins have enabled the visualization of stem cells over extensive periods of time in vivo and ex vivo. The generation of genetically encoded fluorescent protein reporters that are fused with subcellularly localized proteins, such as human histone H2B, has made it possible to direct fluorescent protein reporters to specific subcellular structures and identify single cells in complex populations. This facilitates the visualization of cellular behaviors such as division, movement, and apoptosis at a single-cell resolution and, in principle, allows the prospective and retrospective tracking towards determining the lineage of each cell.

Deuteration distribution estimation with improved sequence coverage for HX/MS experiments.

MOTIVATION: Time-resolved hydrogen exchange (HX) followed by mass spectrometry (MS) is a key technology for studying protein structure, dynamics and interactions. HX experiments deliver a time-dependent distribution of deuteration levels of peptide sequences of the protein of interest. The robust and complete estimation of this distribution for as many peptide fragments as possible is instrumental to understanding dynamic protein-level HX behavior. Currently, this data interpretation step still is a bottleneck in the overall HX/MS workflow. RESULTS: We propose HeXicon, a novel algorithmic workflow for automatic deuteration distribution estimation at increased sequence coverage. Based on an L(1)-regularized feature extraction routine, HeXicon extracts the full deuteration distribution, which allows insight into possible bimodal exchange behavior of proteins, rather than just an average deuteration for each time point. Further, it is capable of addressing ill-posed estimation problems, yielding sparse and physically reasonable results. HeXicon makes use of existing peptide sequence information, which is augmented by an inferred list of peptide candidates derived from a known protein sequence. In conjunction with a supervised classification procedure that balances sensitivity and specificity, HeXicon can deliver results with increased sequence coverage. AVAILABILITY: The entire HeXicon workflow has been implemented in C++ and includes a graphical user interface. It is available at http://hci.iwr.uni-heidelberg.de/software.php. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.