A rapid, high-throughput, viral infectivity assay using automated brightfield microscopy with machine learning.

Infectivity assays are essential for the development of viral vaccines, antiviral therapies, and the manufacture of biologicals. Traditionally, these assays take 2-7 days and require several manual processing steps after infection. We describe an automated viral infectivity assay (AVIATM), using convolutional neural networks (CNNs) and high-throughput brightfield microscopy on 96-well plates that can quantify infection phenotypes within hours, before they are manually visible, and without sample preparation. CNN models were trained on HIV, influenza A virus, coronavirus 229E, vaccinia viruses, poliovirus, and adenoviruses, which together span the four major categories of virus (DNA, RNA, enveloped, and non-enveloped). A sigmoidal function, fit between virus dilution curves and CNN predictions, results in sensitivity ranges comparable to or better than conventional plaque or TCID50 assays, and a precision of ∼10%, which is considerably better than conventional infectivity assays. Because this technology is based on sensitizing CNNs to specific phenotypes of infection, it has potential as a rapid, broad-spectrum tool for virus characterization, and potentially identification.

Clinical Impact and Generalizability of a Computer-Assisted Diagnostic Tool to Risk-Stratify Lung Nodules With CT.

OBJECTIVE: To evaluate whether an imaging classifier for radiology practice can improve lung nodule classification and follow-up. METHODS: A machine learning classifier was developed and trained using imaging data from the National Lung Screening Trial (NSLT) to produce a malignancy risk score (malignancy Similarity Index [mSI]) for individual lung nodules. In addition to NLST cohorts, external cohorts were developed from a tertiary referral lung cancer screening program data set and an external nonscreening data set of all nodules detected on CT. Performance of the mSI combined with Lung-RADS was compared with Lung-RADS alone and the Mayo and Brock risk calculators. RESULTS: We analyzed 963 subjects and 1,331 nodules across these cohorts. The mSI was comparable in accuracy (area under the curve = 0.89) to existing clinical risk models (area under the curve = 0.86-0.88) and independently predictive in the NLST cohort of 704 nodules. When compared with Lung-RADS, the mSI significantly increased sensitivity across all cohorts (25%-117%), with significant increases in specificity in the screening cohorts (17%-33%). When used in conjunction with Lung-RADS, use of mSI would result in earlier diagnoses and reduced follow-up across cohorts, including the potential for early diagnosis in 42% of malignant NLST nodules from prior-year CT scans. CONCLUSION: A computer-assisted diagnosis software improved risk classification from chest CTs of screening and incidentally detected lung nodules compared with Lung-RADS. mSI added predictive value independent of existing radiological and clinical variables. These results suggest the generalizability and potential clinical impact of a tool that is straightforward to implement in practice.

Ribosomal protein L5 facilitates rDNA-bundled condensate and nucleolar assembly.

The nucleolus is the site of ribosome assembly and formed through liquid-liquid phase separation. Multiple ribosomal DNA (rDNA) arrays are bundled in the nucleolus, but the underlying mechanism and significance are unknown. In the present study, we performed high-content screening followed by image profiling with the wndchrm machine learning algorithm. We revealed that cells lacking a specific 60S ribosomal protein set exhibited common nucleolar disintegration. The depletion of RPL5 (also known as uL18), the liquid-liquid phase separation facilitator, was most effective, and resulted in an enlarged and un-separated sub-nucleolar compartment. Single-molecule tracking analysis revealed less-constrained mobility of its components. rDNA arrays were also unbundled. These results were recapitulated by a coarse-grained molecular dynamics model. Transcription and processing of ribosomal RNA were repressed in these aberrant nucleoli. Consistently, the nucleoli were disordered in peripheral blood cells from a Diamond-Blackfan anemia patient harboring a heterozygous, large deletion in RPL5 Our combinatorial analyses newly define the role of RPL5 in rDNA array bundling and the biophysical properties of the nucleolus, which may contribute to the etiology of ribosomopathy.

Predicting physiological aging rates from a range of quantitative traits using machine learning.

It is widely thought that individuals age at different rates. A method that measures "physiological age" or physiological aging rate independent of chronological age could therefore help elucidate mechanisms of aging and inform an individual's risk of morbidity and mortality. Here we present machine learning frameworks for inferring individual physiological age from a broad range of biochemical and physiological traits including blood phenotypes (e.g., high-density lipoprotein), cardiovascular functions (e.g., pulse wave velocity) and psychological traits (e.g., neuroticism) as main groups in two population cohorts SardiNIA (~6,100 participants) and InCHIANTI (~1,400 participants). The inferred physiological age was highly correlated with chronological age (R2 > 0.8). We further defined an individual's physiological aging rate (PAR) as the ratio of the predicted physiological age to the chronological age. Notably, PAR was a significant predictor of survival, indicating an effect of aging rate on mortality. Our trait-based PAR was correlated with DNA methylation-based epigenetic aging score (r = 0.6), suggesting that both scores capture a common aging process. PAR was also substantially heritable (h2~0.3), and a subsequent genome-wide association study of PAR identified significant associations with two genetic loci, one of which is implicated in telomerase activity. Our findings support PAR as a proxy for an underlying whole-body aging mechanism. PAR may thus be useful to evaluate the efficacy of treatments that target aging-related deficits and controllable epidemiological factors.

Radiomics in the evaluation of lung nodules: Intrapatient concordance between full-dose and ultra-low-dose chest computed tomography.

PURPOSE: The purpose of this study was to retrospectively evaluate the quantitative and qualitative intrapatient concordance of pulmonary nodule risk assessment by commercially available radiomics software between full-dose (FD) chest-CT and ultra-low-dose (ULD) chest CT. MATERIALS AND METHODS: Between July 2013 and September 2015, 68 patients (52 men and16 women; mean age, 65.5±10.6 [SD] years; range: 35-87 years) with lung nodules≥5mm and<30mm who underwent the same day FD chest CT (helical acquisition; 120kV; automated tube current modulation) and ULD chest CT (helical acquisition; 135kV; 10mA fixed) were retrospectively included. Each nodule on each acquisition was assessed by a commercial radiomics software providing a similarity malignancy index (mSI), classifying it as "benign-like" (mSI<0.1); "malignant-like" (mSI>0.9) or "undetermined" (0.1≤mSI≤0.9). Intrapatient qualitative agreement was evaluated with weighted Cohen-Kappa test and quantitative agreement with intraclass correlation coefficient (ICC). RESULTS: Ninety-nine lung nodules with a mean size of 9.14±4.3 (SD) mm (range: 5-25mm) in 68 patients (mean 1.46 nodule per patient; range: 1-5) were assessed; mean mSI was 0.429±0.331 (SD) (range: 0.001-1) with FD chest CT (22/99 [22%] "benign-like", 67/99 [68%] "undetermined" and 10/99 [10%] "malignant-like") and mean mSI was 0.487±0.344 (SD) (range: 0.002-1) with ULD chest CT (20/99 [20%] "benign-like", 59/99 [60%] "undetermined" and 20/99 [20%] "malignant-like"). Qualitative and quantitative agreement of FD chest CT with ULD chest CT were "good" with Kappa value of 0.60 (95% CI: 0.46-0.74) and ICC of 0.82 (95% CI: 0.73-0.87), respectively. CONCLUSION: A good agreement in malignancy similarity index can be obtained between ULD chest CT and FD chest CT using radiomics software. However, further studies must be done with more case material to confirm our results and elucidate the diagnostic capabilities of radiomics software using ULD chest CT for lung nodule characterization by comparison with FD chest CT.

Computational analysis of morphological and molecular features in gastric cancer tissues.

Biological morphologies of cells and tissues represent their physiological and pathological conditions. The importance of quantitative assessment of morphological information has been highly recognized in clinical diagnosis and therapeutic strategies. In this study, we used a supervised machine learning algorithm wndchrm to classify hematoxylin and eosin (H&E)-stained images of human gastric cancer tissues. This analysis distinguished between noncancer and cancer tissues with different histological grades. We then classified the H&E-stained images by expression levels of cancer-associated nuclear ATF7IP/MCAF1 and membranous PD-L1 proteins using immunohistochemistry of serial sections. Interestingly, classes with low and high expressions of each protein exhibited significant morphological dissimilarity in H&E images. These results indicated that morphological features in cancer tissues are correlated with expression of specific cancer-associated proteins, suggesting the usefulness of biomolecular-based morphological classification.

Age-related changes of the retinal microvasculature.

PURPOSE: Blood vessels of the retina provide an easily-accessible, representative window into the condition of microvasculature. We investigated how retinal vessel structure captured in fundus photographs changes with age, and how this may reflect features related to patient health, including blood pressure. RESULTS: We used two approaches. In the first approach, we segmented the retinal vasculature from fundus photographs and then we correlated 25 parameterized aspects ("traits")-comprising 15 measures of tortuosity, 7 fractal ranges of self-similarity, and 3 measures of junction numbers-with participant age and blood pressure. In the second approach, we examined entire fundus photographs with a set of algorithmic CHARM features. We studied 2,280 Sardinians, ages 20-28, and an U.S. based population from the AREDS study in 1,178 participants, ages 59-84. Three traits (relating to tortuosity, vessel bifurcation number, and vessel endpoint number) showed significant changes with age in both cohorts, and one additional trait (relating to fractal number) showed a correlation in the Sardinian cohort only. When using second approach, we found significant correlations of particular CHARM features with age and blood pressure, which were stronger than those detected when using parameterized traits, reflecting a greater signal from the entire photographs than was captured in the segmented microvasculature. CONCLUSIONS: These findings demonstrate that automated quantitative image analysis of fundus images can reveal general measures of patient health status.

Transcriptome States Reflect Imaging of Aging States.

In this study, we describe a morphological biomarker that detects multiple discrete subpopulations (or "age-states") at several chronological ages in a population of nematodes (Caenorhabditis elegans). We determined the frequencies of three healthy adult states and the timing of the transitions between them across the lifespan. We used short-lived and long-lived strains to confirm the general applicability of the state classifier and to monitor state progression. This exploration revealed healthy and unhealthy states, the former being favored in long-lived strains and the latter showing delayed onset. Short-lived strains rapidly transitioned through the putative healthy state. We previously found that age-matched animals in different age-states have distinct transcriptome profiles. We isolated animals at the beginning and end of each identified state and performed microarray analysis (principal component analysis, relative sample to sample distance measurements, and gene set enrichment analysis). In some comparisons, chronologically identical individuals were farther apart than morphologically identical individuals isolated on different days. The age-state biomarker allowed assessment of aging in a novel manner, complementary to chronological age progression. We found hsp70 and some small heat shock protein genes are expressed later in adulthood, consistent with the proteostasis collapse model.

Differential Aging Signals in Abdominal CT Scans.

RATIONALE AND OBJECTIVES: Changes in the composition of body tissues are major aging phenotypes, but they have been difficult to study in depth. Here we describe age-related change in abdominal tissues observable in computed tomography (CT) scans. We used pattern recognition and machine learning to detect and quantify these changes in a model-agnostic fashion. MATERIALS AND METHODS: CT scans of abdominal L4 sections were obtained from Baltimore Longitudinal Study of Aging (BLSA) participants. Age-related change in the constituent tissues were determined by training machine classifiers to differentiate age groups within male and female strata ("Younger" at 50-70 years old vs "Older" at 80-99 years old). The accuracy achieved by the classifiers in differentiating the age cohorts was used as a surrogate measure of the aging signal in the different tissues. RESULTS: The highest accuracy for discriminating age differences was 0.76 and 0.72 for males and females, respectively. The classification accuracy was 0.79 and 0.71 for adipose tissue, 0.70 and 0.68 for soft tissue, and 0.65 and 0.64 for bone. CONCLUSIONS: Using image data from a large sample of well-characterized pool of participants dispersed over a wide age range, we explored age-related differences in gross morphology and texture of abdominal tissues. This technology is advantageous for tracking effects of biological aging and predicting adverse outcomes when compared to the traditional use of specific molecular biomarkers. Application of pattern recognition and machine learning as a tool for analyzing medical images may provide much needed insight into tissue changes occurring with aging and, further, connect these changes with their metabolic and functional consequences.

Tomatidine enhances lifespan and healthspan in C. elegans through mitophagy induction via the SKN-1/Nrf2 pathway.

Aging is a major international concern that brings formidable socioeconomic and healthcare challenges. Small molecules capable of improving the health of older individuals are being explored. Small molecules that enhance cellular stress resistance are a promising avenue to alleviate declines seen in human aging. Tomatidine, a natural compound abundant in unripe tomatoes, inhibits age-related skeletal muscle atrophy in mice. Here we show that tomatidine extends lifespan and healthspan in C. elegans, an animal model of aging which shares many major longevity pathways with mammals. Tomatidine improves many C. elegans behaviors related to healthspan and muscle health, including increased pharyngeal pumping, swimming movement, and reduced percentage of severely damaged muscle cells. Microarray, imaging, and behavioral analyses reveal that tomatidine maintains mitochondrial homeostasis by modulating mitochondrial biogenesis and PINK-1/DCT-1-dependent mitophagy. Mechanistically, tomatidine induces mitochondrial hormesis by mildly inducing ROS production, which in turn activates the SKN-1/Nrf2 pathway and possibly other cellular antioxidant response pathways, followed by increased mitophagy. This mechanism occurs in C. elegans, primary rat neurons, and human cells. Our data suggest that tomatidine may delay some physiological aspects of aging, and points to new approaches for pharmacological interventions for diseases of aging.

Predicting early symptomatic osteoarthritis in the human knee using machine learning classification of magnetic resonance images from the osteoarthritis initiative.

The purpose of this study is to evaluate the ability of a machine learning algorithm to classify in vivo magnetic resonance images (MRI) of human articular cartilage for development of osteoarthritis (OA). Sixty-eight subjects were selected from the osteoarthritis initiative (OAI) control and incidence cohorts. Progression to clinical OA was defined by the development of symptoms as quantified by the Western Ontario and McMaster Universities Arthritis (WOMAC) questionnaire 3 years after baseline evaluation. Multi-slice T2 -weighted knee images, obtained through the OAI, of these subjects were registered using a nonlinear image registration algorithm. T2 maps of cartilage from the central weight bearing slices of the medial femoral condyle were derived from the registered images using the multiple available echo times and were classified for "progression to symptomatic OA" using the machine learning tool, weighted neighbor distance using compound hierarchy of algorithms representing morphology (WND-CHRM). WND-CHRM classified the isolated T2 maps for the progression to symptomatic OA with 75% accuracy. CLINICAL SIGNIFICANCE: Machine learning algorithms applied to T2 maps have the potential to provide important prognostic information for the development of OA. © 2017 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 35:2243-2250, 2017.

Cockayne syndrome group A and B proteins converge on transcription-linked resolution of non-B DNA.

Cockayne syndrome is a neurodegenerative accelerated aging disorder caused by mutations in the CSA or CSB genes. Although the pathogenesis of Cockayne syndrome has remained elusive, recent work implicates mitochondrial dysfunction in the disease progression. Here, we present evidence that loss of CSA or CSB in a neuroblastoma cell line converges on mitochondrial dysfunction caused by defects in ribosomal DNA transcription and activation of the DNA damage sensor poly-ADP ribose polymerase 1 (PARP1). Indeed, inhibition of ribosomal DNA transcription leads to mitochondrial dysfunction in a number of cell lines. Furthermore, machine-learning algorithms predict that diseases with defects in ribosomal DNA (rDNA) transcription have mitochondrial dysfunction, and, accordingly, this is found when factors involved in rDNA transcription are knocked down. Mechanistically, loss of CSA or CSB leads to polymerase stalling at non-B DNA in a neuroblastoma cell line, in particular at G-quadruplex structures, and recombinant CSB can melt G-quadruplex structures. Indeed, stabilization of G-quadruplex structures activates PARP1 and leads to accelerated aging in Caenorhabditis elegans In conclusion, this work supports a role for impaired ribosomal DNA transcription in Cockayne syndrome and suggests that transcription-coupled resolution of secondary structures may be a mechanism to repress spurious activation of a DNA damage response.

Loss of the integral nuclear envelope protein SUN1 induces alteration of nucleoli.

A supervised machine learning algorithm, which is qualified for image classification and analyzing similarities, is based on multiple discriminative morphological features that are automatically assembled during the learning processes. The algorithm is suitable for population-based analysis of images of biological materials that are generally complex and heterogeneous. Here we used the algorithm wndchrm to quantify the effects on nucleolar morphology of the loss of the components of nuclear envelope in a human mammary epithelial cell line. The linker of nucleoskeleton and cytoskeleton (LINC) complex, an assembly of nuclear envelope proteins comprising mainly members of the SUN and nesprin families, connects the nuclear lamina and cytoskeletal filaments. The components of the LINC complex are markedly deficient in breast cancer tissues. We found that a reduction in the levels of SUN1, SUN2, and lamin A/C led to significant changes in morphologies that were computationally classified using wndchrm with approximately 100% accuracy. In particular, depletion of SUN1 caused nucleolar hypertrophy and reduced rRNA synthesis. Further, wndchrm revealed a consistent negative correlation between SUN1 expression and the size of nucleoli in human breast cancer tissues. Our unbiased morphological quantitation strategies using wndchrm revealed an unexpected link between the components of the LINC complex and the morphologies of nucleoli that serves as an indicator of the malignant phenotype of breast cancer cells.

Biomedical imaging ontologies: A survey and proposal for future work.

BACKGROUND: Ontology is one strategy for promoting interoperability of heterogeneous data through consistent tagging. An ontology is a controlled structured vocabulary consisting of general terms (such as "cell" or "image" or "tissue" or "microscope") that form the basis for such tagging. These terms are designed to represent the types of entities in the domain of reality that the ontology has been devised to capture; the terms are provided with logical definitions thereby also supporting reasoning over the tagged data. AIM: This paper provides a survey of the biomedical imaging ontologies that have been developed thus far. It outlines the challenges, particularly faced by ontologies in the fields of histopathological imaging and image analysis, and suggests a strategy for addressing these challenges in the example domain of quantitative histopathology imaging. RESULTS AND CONCLUSIONS: The ultimate goal is to support the multiscale understanding of disease that comes from using interoperable ontologies to integrate imaging data with clinical and genomics data.

The actin family protein ARP6 contributes to the structure and the function of the nucleolus.

The actin family members, consisting of actin and actin-related proteins (ARPs), are essential components of chromatin remodeling complexes. ARP6, one of the nuclear ARPs, is part of the Snf-2-related CREB-binding protein activator protein (SRCAP) chromatin remodeling complex, which promotes the deposition of the histone variant H2A.Z into the chromatin. In this study, we showed that ARP6 influences the structure and the function of the nucleolus. ARP6 is localized in the central region of the nucleolus, and its knockdown induced a morphological change in the nucleolus. We also found that in the presence of high concentrations of glucose ARP6 contributed to the maintenance of active ribosomal DNA (rDNA) transcription by placing H2A.Z into the chromatin. In contrast, under starvation, ARP6 was required for cell survival through the repression of rDNA transcription independently of H2A.Z. These findings reveal novel pleiotropic roles for the actin family in nuclear organization and metabolic homeostasis.

A Longitudinal Support Vector Regression for Prediction of ALS Score.

Longitudinal studies play a key role in various fields, including epidemiology, clinical research, and genomic analysis. Currently, the most popular methods in longitudinal data analysis are model-driven regression approaches, which impose strong prior assumptions and are unable to scale to large problems in the manner of machine learning algorithms. In this work, we propose a novel longitudinal support vector regression (LSVR) algorithm that not only takes the advantage of one of the most popular machine learning methods, but also is able to model the temporal nature of longitudinal data by taking into account observational dependence within subjects. We test LSVR on publicly available data from the DREAM-Phil Bowen ALS Prediction Prize4Life challenge. Results suggest that LSVR is at a minimum competitive with favored machine learning methods and is able to outperform those methods in predicting ALS score one month in advance.

An open data ecosystem for cell migration research.

Cell migration research has recently become both a high content and a high throughput field thanks to technological, computational, and methodological advances. Simultaneously, however, urgent bioinformatics needs regarding data management, standardization, and dissemination have emerged. To address these concerns, we propose to establish an open data ecosystem for cell migration research.

Computational image analysis of colony and nuclear morphology to evaluate human induced pluripotent stem cells.

Non-invasive evaluation of cell reprogramming by advanced image analysis is required to maintain the quality of cells intended for regenerative medicine. Here, we constructed living and unlabelled colony image libraries of various human induced pluripotent stem cell (iPSC) lines for supervised machine learning pattern recognition to accurately distinguish bona fide iPSCs from improperly reprogrammed cells. Furthermore, we found that image features for efficient discrimination reside in cellular components. In fact, extensive analysis of nuclear morphologies revealed dynamic and characteristic signatures, including the linear form of the promyelocytic leukaemia (PML)-defined structure in iPSCs, which was reversed to a regular sphere upon differentiation. Our data revealed that iPSCs have a markedly different overall nuclear architecture that may contribute to highly accurate discrimination based on the cell reprogramming status.

Genome-wide modeling of complex phenotypes in Caenorhabditis elegans and Drosophila melanogaster.

BACKGROUND: The genetic and molecular basis for many intermediate and end stage phenotypes in model systems such as C. elegans and D. melanogaster has long been known to involve pleiotropic effects and complex multigenic interactions. Gene sets are groups of genes that contribute to multiple biological or molecular phenomena. They have been used in the analysis of large molecular datasets such as microarray data, Next Generation sequencing, and other genomic datasets to reveal pleiotropic and multigenic contributions to phenotypic outcomes. Many model systems lack species specific organized phenotype based gene sets to enable high throughput analysis of large molecular datasets. RESULTS AND DISCUSSION: Here, we describe two novel collections of gene sets in C. elegans and D. melanogaster that are based exclusively on genetically determined phenotypes and use a controlled phenotypic ontology. We use these collections to build genome-wide models of thousands of defined phenotypes in both model species. In addition, we demonstrate the utility of these gene sets in systems analysis and in analysis of gene expression-based molecular datasets and show how they are useful in analysis of genomic datasets connecting multigenic gene inputs to complex phenotypes. CONCLUSIONS: Phenotypic based gene sets in both C. elegans and D. melanogaster are developed, characterized, and shown to be useful in the analysis of large scale species-specific genomic datasets. These phenotypic gene set collections will contribute to the understanding of complex phenotypic outcomes in these model systems.

Systematic repression of transcription factors reveals limited patterns of gene expression changes in ES cells.

Networks of transcription factors (TFs) are thought to determine and maintain the identity of cells. Here we systematically repressed each of 100 TFs with shRNA and carried out global gene expression profiling in mouse embryonic stem (ES) cells. Unexpectedly, only the repression of a handful of TFs significantly affected transcriptomes, which changed in two directions/trajectories: one trajectory by the repression of either Pou5f1 or Sox2; the other trajectory by the repression of either Esrrb, Sall4, Nanog, or Tcfap4. The data suggest that the trajectories of gene expression change are already preconfigured by the gene regulatory network and roughly correspond to extraembryonic and embryonic fates of cell differentiation, respectively. These data also indicate the robustness of the pluripotency gene network, as the transient repression of most TFs did not alter the transcriptomes.