Kyle W. Karhohs's research while affiliated with Broad Institute of MIT and Harvard and other places

Publications (37)

Article
Morphological and gene expression profiling can cost-effectively capture thousands of features in thousands of samples across perturbations by disease, mutation, or drug treatments, but it is unclear to what extent the two modalities capture overlapping versus complementary information. Here, using both the L1000 and Cell Painting assays to profile...
Preprint
Full-text available
Resolving fundamental molecular and functional processes underlying human synaptic development is crucial for understanding normal brain function as well as dysfunction in disease. Based upon increasing evidence of species divergent features of brain cell types, coupled with emerging studies of complex human disease genetics, we developed the first...
Article
Patient stem cell-derived models enable imaging of complex disease phenotypes and the development of scalable drug discovery platforms. Current preclinical methods for assessing cellular activity do not, however, capture the full intricacies of disease-induced disturbances, and instead typically focus on a single parameter, which impairs both the u...
Preprint
Full-text available
Deep profiling of cell states can provide a broad picture of biological changes that occur in disease, mutation, or in response to drug or chemical treatments. Morphological and gene expression profiling, for example, can cost-effectively capture thousands of features in thousands of samples across perturbations, but it is unclear to what extent th...
Preprint
Full-text available
Patient stem cell-derived models enable imaging of complex disease phenotypes and the development of scalable drug discovery platforms. Current preclinical methods for assessing cellular activity do not, however, capture the full intricacies of disease-induced disturbances, and instead typically focus on a single parameter, which impairs both the u...
Article
Full-text available
Background: Identified as an Alzheimer's disease (AD) susceptibility gene by genome wide-association studies, BIN1 has 10 isoforms that are expressed in the Central Nervous System (CNS). The distribution of these isoforms in different cell types, as well as their role in AD pathology still remains unclear. Methods: Utilizing antibodies targeting...
Article
Full-text available
Cellular responses to stimuli can evolve over time, resulting in distinct early and late phases in response to a single signal. DNA damage induces a complex response that is largely orchestrated by the transcription factor p53, whose dynamics influence whether a damaged cell will arrest and repair the damage or will initiate cell death. How p53 res...
Article
Background The mechanisms that regulate platelet biogenesis remain unclear; factors that trigger megakaryocytes (MKs) to initiate platelet production are poorly understood. Platelet formation begins with proplatelets which are cellular extensions originating from the MK cell body. Objectives Proplatelet formation is an asynchronous and dynamic pro...
Article
Full-text available
An amendment to this paper has been published and can be accessed via a link at the top of the paper.
Article
Full-text available
Kidney stones and ureteral stents can cause ureteral colic and pain. By decreasing contractions in the ureter, clinically prescribed oral vasodilators may improve spontaneous stone passage rates and reduce the pain caused by ureteral stenting. We hypothesized that ureteral relaxation can be improved via the local administration of vasodilators and...
Article
Full-text available
Segmenting the nuclei of cells in microscopy images is often the first step in the quantitative analysis of imaging data for biological and biomedical applications. Many bioimage analysis tools can segment nuclei in images but need to be selected and configured for every experiment. The 2018 Data Science Bowl attracted 3,891 teams worldwide to make...
Article
Full-text available
Dysregulated axonal trafficking of mitochondria is linked to neurodegenerative disorders. We report a high-content screen for small-molecule regulators of the axonal transport of mitochondria. Six compounds enhanced mitochondrial transport in the sub-micromolar range, acting via three cellular targets: F-actin, Tripeptidyl peptidase 1 (TPP1), or Au...
Article
Full-text available
Identifying nuclei is often a critical first step in analyzing microscopy images of cells and classical image processing algorithms are most commonly used for this task. Recent developments in deep learning can yield superior accuracy, but typical evaluation metrics for nucleus segmentation do not satisfactorily capture error modes that are relevan...
Article
Biological signals need to be robust and filter small fluctuations yet maintain sensitivity to signals across a wide range of magnitudes. Here, we studied how fluctuations in DNA damage signaling relate to maintenance of long-term cell-cycle arrest. Using live-cell imaging, we quantified division profiles of individual human cells in the course of...
Data
Segmentation steps for the analysis of synthetic images depicting HL60 cell line nuclei. Images are available from the Broad Bioimage Benchmark Collection (https://data.broadinstitute.org/bbbc/BBBC024/), as in Fig 2C of the main paper. Synthetic images with 75% clustering probability and low SNR were chosen for analysis. The data set was generated...
Data
Modules with support for 3D in CellProfiler. Overview of modules available in CellProfiler 3.1.0 for 3D image analysis. 3D, three-dimensional. (TIF)
Data
Segmentation steps for the analysis of synthetic images from the cell tracking challenge (http://www.celltrackingchallenge.net) depicting HL60 cell line nuclei. Images are taken from the Broad Bioimage Benchmark Collection (https://data.broadinstitute.org/bbbc/BBBC035/), as in Fig 2D of the main paper. (A) Original 3D image of HL60 nuclei prior to...
Data
Accuracy of nuclear segmentation using CellProfiler and MorphoLibJ. The fraction of nuclei correctly identified relative to their ground truth was assessed for both CellProfiler (solid line) and MorphoLibJ (dashed line) for the results shown in Fig 1 and S2–S5 Figs. A nucleus was considered correctly segmented at a given threshold if the intersecti...
Data
Comparison of runtimes between CellProfiler 3.0 and CellProfiler 2.2. (Toward the left: CellProfiler 3.0 is faster; toward the right: CellProfiler 2.2 is faster). We compared the performance of example pipelines (available at https://github.com/CellProfiler/examples and http://cellprofiler.org) for CellProfiler 2.2 and 3.0 on OS X 10.12.6 (2.8 GHz...
Data
Distributed-CellProfiler enables processing thousands of images in parallel. A data set of seventeen 384-well plates was processed using Distributed-CellProfiler on an AWS cluster. Each plate comprised 3,456 five-channel images (2,160 × 2,160 pixels). A CellProfiler pipeline was run on each image to identify cells and then extract measurements per...
Data
CellProfiler measurements of mouse blastocysts. Measurements of mouse blastocyst cells and GAPDH transcript foci created by CellProfiler; these serve as the underlying data for Fig 3F. GAPDH, glyceraldehyde 3-phosphate dehydrogenase. (XLSX)
Data
Fiji macros used for each file. Macro code constructed in the MorphoLibJ plugin for Fiji software that was used for analysis of 3D image examples presented in Fig 1 and S2–S5 Figs. (XLSX)
Data
Segmentation steps for the analysis of mouse embryo blastocyst nuclei stained with Hoechst. Images are available from the Broad Bioimage Benchmark Collection (https://data.broadinstitute.org/bbbc/BBBC032/), as in Fig 2A of the main paper. (A) Original 3D image of blastocyst nuclei prior to analysis. (B) Evaluation of CellProfiler 3.0 performance in...
Data
Image segmentation speed comparison of CellProfiler and the MorphoLibJ 1.3.3 plugin in Fiji (ImageJ 1.51n) on x-64 PC (Dell 7280) Windows 10 Pro (2.8 GHz Intel Core i7-7600U CPU and 16 GB 2133 MHz DDR4 SDRAM). Units of time are in seconds. One image representing the 72-hour time point (72 hours) was used to compare CellProfiler 3.0 and Fiji present...
Data
Example pipeline and nuclei image for deep learning in CellProfiler 3.0. Example pipeline and image needed to run the ClassifyPixels-Unet module. (ZIP)
Data
CellProfiler measurements of hiPSCs. Per cell measurements of hiPSCs created by CellProfiler; these serve as the underlying data for Fig 4D. hiPSC, human induced pluripotent stem cell. (XLSX)
Data
Example pipelines and configuration files needed to run CellProfiler on AWS. These pipelines and files can be used to run Distributed-CellProfiler; they are specifically configured to run the Cell Painting assay [7]. AWS, Amazon Web Services. (ZIP)
Article
Full-text available
Author summary The “big-data revolution” has struck biology: it is now common for robots to prepare cell samples and take thousands of microscopy images. Looking at the resulting images by eye would be extremely tedious, not to mention subjective. Thus, many biologists find they need software to analyze images easily and accurately. The third major...
Data
Segmentation of U2OS cells in images, using the deep learning based ClassifyPixels-Unet plugin. Image is available from the Broad Bioimage Benchmark Collection (https://data.broadinstitute.org/bbbc/BBBC022/, filename XMtest_B12_s2_w19F7E0279-D087-4B5E-9899-61971C29CB78.tif, see S4 File). The U-Net model was trained using 150 manually annotated DAPI...
Data
CellProfiler 3D module run times. CPU and wall-clock runtimes for individual CellProfiler modules in the example pipelines; the summary of this data can be seen in S2 Table. 3D, three-dimensional; CPU, central processing unit (XLSX)
Data
Segmentation steps for the analysis of mouse trophoblast stem cell nuclei stained with Hoechst. Images are available from the Broad Bioimage Benchmark Collection (https://data.broadinstitute.org/bbbc/BBBC033/), as in Fig 2B of the main paper. (A) Original 3D stem cell nuclei image prior to analysis. (B) Evaluation of CellProfiler 3.0 performance in...
Data
Code to asses per-object segmentation accuracy in CellProfiler and MorphoLibJ. A file to reproduce the results presented in S6 Fig. This contains the ground truth images, CellProfiler-produced segmentations, and MorphoLibJ-produced segmentations, as well as a Jupyter notebook that can be run from the directory once unzipped to replicate the code. (...
Preprint
Full-text available
Identifying nuclei is often a critical first step in analyzing microscopy images of cells, and classical image processing algorithms are still commonly used for this task. Recent studies indicate that deep learning may yield superior accuracy, but its performance has not been evaluated for high-throughput nucleus segmentation in large collections o...
Poster
Human motor neurons (MN) derived from induced pluripotent stem cells (iPS) of familial ALS patients with C9orf72, SOD1 and FUS mutations exhibit in a manner similar to the patients, a hyperexcitability phenotype that contributes to susceptibility to cell death. The exact mechanisms leading from these diverse mutations to hyperexcitability are still...
Article
Cells transmit information through molecular signals that often show complex dynamical patterns. The dynamic behavior of the tumor suppressor p53 varies depending on the stimulus; in response to double-strand DNA breaks, it shows a series of repeated pulses. Using a computational model, we identified a sequence of precisely timed drug additions tha...

Citations

... Then, the feature correlation distribution of replicate wells (percent replicating) or MOA matching wells (percent matching) is calculated. The percentage of the replicate or matching distribution that is above the 95th percentile of the null distribution is the percent replicating or percent matching score, respectively [16]. ...
... Patient-specific iPSC-derived neuronal cells can be used for assessing cell-based traits associated with disease-related mutations. Previously, an open-source cloud-based image processing and analysis platform (CELLXPEDITE) was used to screen compounds that revert the multiparametric disease profile of ALS back to that of healthy motor neurons [235]. This is advantageous when compared to the unidimensional characterisation of other screening methods for diseases such as ALS [236], as it allows for a simultaneous measurement of multiple parameters that would be more reflective of a complex disease. ...
... This equated to an accuracy of 0.52 (random accuracy = 0.006). To assess the influence of well-location effects which are common in microscopy 38,39 , we compared the model's performance on edge wells versus non-edge wells. We found minor differences in classification accuracy (0.54 vs .50) ...
... The analysis in this work is based on two datasets of individual human cells that were published in [23,29]. The cells were monitored using fluorescence live-cell imaging for about a week after being subjected to irradiation. ...
... Unrepaired ICLs lead to a complete block of the separation of the complementary DNA strands and therefore inhibit several essential processes such as DNA replication and transcription [28,30]. If the replication fork stalls at an ICL, it is processed into a double-strand break (DSB), which in turn stops cell cycle progression [28,[31][32][33]. Another consequence of ICLs is that DNA-protein interactions (e.g., transcription factors binding to DNA) at the affected sites of the DNA are no longer possible. ...
... One identified IBA1-positive cells expressing BIN1 in immunohistochemical analysis of post-mortem brain tissue from patients with AD [8]. Another study reported detection of BIN1 isoforms 12 and 6 in the nucleus of CD45 + microglia using antibodies raised against BIN1 exons 11 and 13; notably, the major microglial BIN1 isoform 10 lacking both exons 11 and 13 was not observed by immunostaining in the previous study [65]. In this regard, it is worth mentioning that exon 11, which codes for a polybasic sequence that confers binding to phosphoinositides and is essential for BIN1-induced membrane tubulation [66], is spliced out from Bin1 transcripts in adult mouse brain microglia and human induced iPSCs-derived microglia (Fig. S2A). ...
... 4 In our recent work, we have focused on cell surface area, perimeter, circularity (FormFactor), pseudopodia (endpoint) number, and total pseudopodia length. 1 Examples of the processing steps from which these outputs are measured are seen in Figure 2. Cell morphology characteristics have been shown to change under various conditions, such as chronic stress. 1 Protocols similar to this have also been used in other biological contexts and cell types, demonstrating the broad applicability of quantifying cell morphology. 8,9 Cell morphology phenotypes can be indicative of organism characteristics; for example, differences in cell morphology have been associated with donor age more closely than other established biomolecular measurements such as DNA damage response or adenosine triphosphate content. 10 ...
... Data Science Bowl 2018(DSB2018) [58] includes 670 artificially segmented nuclei images from experiments across different tissue samples, staining protocols, imaging conditions and microscopes. This dataset consists of 29455 manually annotated nuclei boundaries by a team of expert biologists. ...
... Given that ureteral smooth muscle cells (USMCs) cannot be obtained commercially, we cultured USMCs from rats and patient ureters using a de novo technique [14]. Briefly, the ureters harvested from patients and rats were immediately placed in cold Krebs solution, cut into 5 mm sections, and incubated on a plate containing trypsin. ...
... The 2018 data science bowl (DSB) dataset was created as a challenge for generic segmentation of nuclei of cells in a diverse set of stained two-dimensional (2D) microscopic images [9]. The training set contains 670 images from both bright-stained and fluorescence modalities of microscopic images with sizes 256 Â 256 Â 3. In addition to the images captured under various lighting conditions, corresponding annotations (segmentation masks) for each image are also provided to be used as ground truth. ...