Search

Many biological macromolecules can phase-separate in the cell and form highly concentrated condensates. The mesoscopic dynamics of these assemblies have been widely characterized, but their behavior at the molecular scale has remained more elusive. Here we investigate condensates of two highly charged disordered human proteins as a characteristic example of liquid-liquid phase separation. The dense phase is 1000 times more concentrated and has 300 times higher bulk viscosity than the dilute phase. However, single-molecule spectroscopy in individual droplets reveals that the polypeptide chains are remarkably dynamic, with sub-microsecond reconfiguration times. We rationalize this behavior with large-scale all-atom molecular-dynamics simulations, which reveal an unexpectedly similar short-range molecular environment in the dense and dilute phases, suggesting that local biochemical processes and interactions can remain exceedingly rapid in phase-separated systems.

The recognition of intrinsically disordered proteins (IDPs) is highly dependent on dynamics owing to the lack of structure. Here we studied the interplay between dynamics and molecular recognition in IDPs with a combination of time‐resolving tools on timescales ranging from femtoseconds to nanoseconds. We interrogated conformational dynamics and surface water dynamics and its attenuation upon partner binding using two IDPs, IBB and Nup153FG, both of central relevance to the nucleocytoplasmic transport machinery. These proteins bind the same nuclear transport receptor (Importinβ) with drastically different binding mechanisms, coupled folding–binding and fuzzy complex formation, respectively. Solvent fluctuations in the dynamic interface of the Nup153FG‐Importinβ fuzzy complex were largely unperturbed and slightly accelerated relative to the unbound state. In the IBB‐Importinβ complex, on the other hand, substantial relative slowdown of water dynamics was seen in a more rigid interface. These results show a correlation between interfacial water dynamics and the plasticity of IDP complexes, implicating functional relevance for such differential modulation in cellular processes, including nuclear transport.

Modern deep learning systems for medical image classification have demonstrated exceptional capabilities for distinguishing between image based medical categories. However, they are severely hindered by their ina-bility to explain the reasoning behind their decision making. This is partly due to the uninterpretable continuous latent representations of neural net-works. Emergent languages (EL) have recently been shown to enhance the capabilities of neural networks by equipping them with symbolic represen-tations in the framework of referential games. Symbolic representations are one of the cornerstones of highly explainable good old fashioned AI (GOFAI) systems. In this work, we demonstrate for the first time, the emer-gence of deep symbolic representations of emergent language in the frame-work of image classification. We show that EL based classification models can perform as well as, if not better than state of the art deep learning mod-els. In addition, they provide a symbolic representation that opens up an entire field of possibilities of interpretable GOFAI methods involving symbol manipulation. We demonstrate the EL classification framework on immune cell marker based cell classification and chest X-ray classification using the CheXpert dataset. Code is available online at https://github.com/AriChow/EL.

The relationship between cellular architecture and cellular state and function is apparent, but not yet completely understood. Precise characterization of cellular state is important in many fields, from pathology to synthetic biology. High-content high-throughput microscopy is now more than ever accessible to researchers. This allows for collection of large amount of cellular images. Naturally, the analysis of this data cannot be left to manual investigation and needs to resort to the use of efficient computing algorithms for cellular detection, segmentation, and tracking. Annotation is required for building high quality algorithms. Medical professionals and researchers spend a lot of effort and time in annotating cells. This task has proved to be very repetitive and time consuming. The experts’ time is valuable and should be used effectively. Our hypothesis is that active deep learning will help to share some of the burden that researchers face in their everyday work. In this paper, we focus specifically on the problem of cellular segmentation.We approach the segmentation task using a classification framework. Each pixel in the image is classified based on whether the patch around it resides on the interior, boundary or exterior of the cell. Deep convolutional neural networks (CNN) are used to perform the classification task. Active learning is the method used to reduce the annotation burden. Uncertainty sampling, a popular active learning framework is used in conjunction with CNN to segment the cells in the image. Three datasets of mammalian nuclei and cytoplasm are used for this work. We show that active deep learning significantly reduces the number of training samples required and also improves the quality of segmentation.

We present a method for automatic cell classification in tissue samples using an automated training set from multiplexed immunofluorescence images. The method utilizes multiple markers stained in situ on a single tissue section on a robust hyperplex immunofluorescence platform (Cell DIVE, GE Healthcare) that provides multi-channel images allowing analysis at single cell/sub-cellular levels. The cell classification method consists of two steps: first, an automated training set from every image is generated using marker-to-cell staining information. This mimics how a pathologist would select samples from a very large cohort at the image level. In the second step, a probability model is inferred from the automated training set. The probabilistic model captures staining patterns in mutually exclusive cell types and builds a single probability model for the data cohort. We have evaluated the proposed approach to classify: i) immune cells in cancer and ii) brain cells in neurological degenerative diseased tissue with average accuracies above 95%., 2019 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

We introduce Symbolic Variational Autoencoders which generate images from symbols that represent semantic concepts. Unlike generic Variational Autoencoders (VAEs) or Generative Adversarial Networks (GANs), the latent distribution from the Symbolic Variational Autoencoder is discrete. The symbols are learned in a completely unsupervised manner by reconstructing images from symbolic encodings. We demonstrate the efficacy of our symbolic approach on the MNIST and FashionMNIST datasets. Results indicate that symbolic encodings naturally form a grammar, where unique strings of symbols map to different semantic concepts. We further explore how changing these symbols affects the final image that is generated.

We present ESCELL, a method for developing an emergent symbolic language of communication between multiple agents reasoning about cells. We show how agents are able to cooperate and communicate successfully in the form of symbols similar to human language to accomplish a task in the form of a referential game (Lewis' signaling game). In one form of the game, a sender and a receiver observe a set of cells from 5 different cell phenotypes. The sender is told one cell is a target and is allowed to send one symbol to the receiver from a fixed arbitrary vocabulary size. The receiver relies on the information in the symbol to identify the target cell. We train the sender and receiver networks to develop an innate emergent language between themselves to accomplish this task. We observe that the networks are able to successfully identify cells from 5 different phenotypes with an accuracy of 93.2%. We also introduce a new form of the signaling game where the sender is shown one image instead of all the images that the receiver sees. The networks successfully develop an emergent language to get an identification accuracy of 77.8%., IEEE International Symposium on Biomedical Imaging (IEEE ISBI 2020)

Recent advances in methods focused on the grounding problem have resulted in techniques that can be used to construct a symbolic language associated with a specific domain. Inspired by how humans communicate complex ideas through language, we developed a generalized Symbolic Semantic (S2) framework for interpretable segmentation. Unlike adversarial models (e.g., GANs), we explicitly model cooperation between two agents, a Sender and a Receiver, that must cooperate to achieve a common goal. The Sender receives information from a high layer of a segmentation network and generates a symbolic sentence derived from a categorical distribution. The Receiver obtains the symbolic sentences and cogenerates the segmentation mask. In order for the model to converge, the Sender and Receiver must learn to communicate using a private language. We apply our architecture to segment tumors in the TCGA dataset. A UNet-like architecture is used to generate input to the Sender network which produces a symbolic sentence, and a Receiver network cogenerates the segmentation mask based on the sentence. Our Sementation framework achieved similar or better performance compared with state-of-the-art segmentation methods. In addition, our results suggest direct interpretation of the symbolic sentences to discriminate between normal and tumor tissue, tumor morphology, and other image characteristics.

We investigate methods of microstructure representation for the purpose of predicting processing condition from microstructure image data. A binary alloy (uranium-molybdenum) that is currently under development as a nuclear fuel was studied for the purpose of developing an improved machine learning approach to image recognition, characterization, and building predictive capabilities linking microstructure to processing conditions. Here, we test different microstructure representations and evaluate model performance based on the F1 score. A F1 score of 95.1% was achieved for distinguishing between micrographs corresponding to ten different thermo-mechanical material processing conditions. We find that our newly developed microstructure representation describes image data well, and the traditional approach of utilizing area fractions of different phases is insufficient for distinguishing between multiple classes using a relatively small, imbalanced original data set of 272 images. To explore the applicability of generative methods for supplementing such limited data sets, generative adversarial networks were trained to generate artificial microstructure images. Two different generative networks were trained and tested to assess performance. Challenges and best practices associated with applying machine learning to limited microstructure image data sets is also discussed. Our work has implications for quantitative microstructure analysis, and development of microstructure-processing relationships in limited data sets typical of metallurgical process design studies., Comment: 14 pages, 15 figures

Recent work on unsupervised learning has explored the feasibility of semantic analysis and interpretation via Emergent Language (EL) models. As EL requires some form of numerical embedding, it remains unclear which type is required in order for the EL to properly capture certain semantic concepts associated with a given task. In this paper, we compare different approaches that can be used to generate such embeddings: unsupervised and supervised. We start by producing a large dataset using a single-agent simulation environment. In these experiments, a purpose-driven agent attempts to accomplish a number of tasks. These tasks are performed in a synthetic cityscape environment, which includes houses, banks, theaters, and restaurants. Given such experiences, specification of the associated goal structure constitutes a narrative. We investigate the feasibility of producing an EL from raw pixel data with the hope that resulting descriptions can be used to infer the underlying narrative structure. Our initial experiments show that a supervised learning approach yields embeddings and EL descriptions that capture narrative structure. Alternatively, an unsupervised learning approach results in greater ambiguity with respect to the narrative.

Background Under oxidative stress cytoplasmic aminoacyl-tRNA synthetase (aaRSs) substrate specificity can be compromised, leading to tRNA mischarging and mistranslation of the proteome. Whether similar processes occur in mitochondria, which are major cellular sources of reactive oxygen species (ROS), is unknown. However, relaxed substrate specificity in yeast mitochondrial phenylalanyl-tRNA synthetase (ScmitPheRS) has been reported to increase tRNA mischarging and blocks mitochondrial biogenesis. Methods Non-reducing denaturing PAGE, cysteine reactivity studies, MALDI-TOF mass spectrometry, enzyme assay, western blot, growth assay, circular dichroism, dynamic light scattering and fluorescence spectroscopy were used to study the effect of oxidative stress on ScmitPheRS activity. Results ScmitPheRS is reversibly inactivated under oxidative stress. The targets for oxidative inactivation are two conserved cysteine residues resulting in reversible intra-molecular disulfide bridge formation. Replacement of either conserved cysteine residue increased viability during growth under oxidative stress. Conclusion Formation of intra-molecular disulfide bridge under oxidative stress hinders the tRNAPhe binding of the enzyme, thus inactivating ScmitPheRS reversibly. General significance The ScmitPheRS activity is compromised under oxidative stress due to formation of intra-molecular disulfide bridge. The sensitivity of ScmitPheRS to oxidation may provide a protective mechanism against error-prone translation under oxidative stress.

Cellular RNAs often colocalize with cytoplasmic, membrane-less ribonucleoprotein (RNP) granules enriched for RNA processing enzymes, termed processing bodies (PBs). Here, we track the dynamic localization of individual miRNAs, mRNAs and long non-coding RNAs (lncRNAs) to PBs using intracellular single-molecule fluorescence microscopy. We find that unused miRNAs stably bind to PBs, whereas functional miRNAs, repressed mRNAs and lncRNAs both transiently and stably localize within either the core or periphery of PBs, albeit to different extents. Consequently, translation potential and 3` versus 5` placement of miRNA target sites significantly impact PB-localization dynamics of mRNAs. Using computational modeling and supporting experimental approaches we show that partitioning into the PB phase attenuates mRNA silencing, suggesting that physiological mRNA turnover occurs predominantly outside PBs. Our data support a PB role instead in sequestering unused miRNAs for surveillance and provides a framework for investigating the dynamic assembly of RNP granules by phase separation at single-molecule resolution.

Computer vision and machine learning methods were applied to the challenge of automatic microstructure recognition. Here, a case study on dendritic morphologies was performed. Two classification tasks were completed, and involved distinguishing between micrographs that depict dendritic morphologies from those that do not contain this particular microstructural feature (Task 1), and from those micrographs identified as depicting dendrites, different cross-sectional views (longitudinal or transverse) were identified (Task 2). Data sets were comprised of images taken over a range of magnifications, from materials with different compositions and varying orientations of microstructural features. Feature extraction and dimensionality reduction were performed prior to training machine learning algorithms to classify microstructural image data. Visual bag of words, texture and shape statistics, and pre-trained convolutional neural networks (deep learning algorithms) were used for feature extraction. Classification was then performed using support vector machine, voting, nearest neighbors, and random forest models. For each model, classification was completed using full (original size) and reduced feature vectors for each feature extraction method tested. Performance comparisons were done to evaluate all possible combinations of feature extraction, selection, and classifiers for the task of micrograph classification. Results demonstrate that pre-trained neural networks represent microstructure image data well, and when used for feature extraction yield the highest classification accuracies for the majority of classifier and feature selection methods tested. Thus, deep learning algorithms can successfully be applied to micrograph recognition tasks. Maximum classification accuracies of 91.85 ± 4.25% and 97.37 ± 3.33% for Tasks 1 and 2 respectively, were achieved. This work is a broad investigation of computer vision and machine learning methods that acts as a step towards applying these established methods to more sophisticated materials recognition or characterization tasks. The approach presented here could offer improvements over established stereological measurements by removing the requirement of expert knowledge (bias) for interpretation of image data prior to characterization.

The lack of computational support has significantly slowed down automatic understanding of endangered languages. In this paper, we take Nyushu (simplified Chinese: 女书; literally: “women’s writing”) as a case study to present the first computational approach that combines Computer Vision and Natural Language Processing techniques to deeply understand an endangered language. We developed an end-to-end system to read a scanned hand-written Nyushu article, segment it into characters, link them to standard characters, and then translate the article into Mandarin Chinese. We propose several novel methods to address the new challenges introduced by noisy input and low resources, including Nyushu-specific feature selection for character segmentation and linking, and character linking lattice based Machine Translation. The end-to-end system performance indicates that the system is a promising approach and can serve as a standard benchmark.

Editors at Science requested our input on the above discussion (comment by Best et al . and response by Riback et al .) because both sets of authors use our data from Fuertes et al . (2017) to support their arguments. The topic of discussion pertains to the discrepant inferences drawn from SAXS versus FRET measurements regarding the dimensions of intrinsically disordered proteins (IDPs) in aqueous solvents. Using SAXS measurements on labeled and unlabeled proteins, we ruled out the labels used for FRET measurements as the cause of discrepant inferences between the two methods. Instead, we propose that FRET and SAXS provide complementary readouts because of a decoupling of size and shape fluctuations that is intrinsic to finite-sized, heteropolymeric IDPs. Accounting for this decoupling resolves the discrepant inferences between the two methods, thus making a case for the utility of both methods.