Bots for Software-Assisted Analysis of Image-Based Transcriptomics

Cicconet, Marcelo; Hochbaum, Daniel; Richmond, David L.; Sabatini, Bernardo L.

doi:10.1101/172296

Cited by 10 publications

(11 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…3. Furthermore, all of our models, even those trained with fewer annotations, show better performances in terms of the NCS than do those (approximately 0.6) reported in [31].…”

Section: B Overall Performancementioning

confidence: 63%

“…We evaluated the proposed method on the nucleus dataset published in [31]. The dataset contains fully annotated DAPI-stained cell nuclei, and we refer to it herein as DH100.…”

Section: Resultsmentioning

confidence: 99%

“…Normalized coverage score (NCS) values were used in [31] to evaluate the matching quality between the ground truth and the predicted segmentation. The NCS is defined as…”

Section: A Evaluationmentioning

confidence: 99%

“…Then, some heavy postprocessing procedures are used to compose or fuse all pixel-level predictions to produce the segmentation of individual nuclei. Here, U-Net is used for pixel classification, and it has been shown that a comparable performance can be achieved when using a stacked random forest classifier as the classification backend [31]. A mask region-based convolutional neural network (Mask R-CNN), on the other hand, is an instance-aware framework and directly operates on a single-nucleus level.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Robust Nucleus Detection With Partially Labeled Exemplars

et al. 2019

View full text Add to dashboard Cite

Quantitative analysis of cell nuclei in microscopic images is an essential yet challenging source of biological and pathological information. The major challenge is accurate detection and segmentation of densely packed nuclei in images acquired under a variety of conditions. Mask R-CNN-based methods have achieved state-of-the-art nucleus segmentation. However, the current pipeline requires fully annotated training images, which are time consuming to create and sometimes noisy. Importantly, nuclei often appear similar within the same image. This similarity could be utilized to segment nuclei with only partially labeled training examples. We propose a simple yet effective region-proposal module for the current Mask R-CNN pipeline to perform few-exemplar learning. To capture the similarities between unlabeled regions and labeled nuclei, we apply decomposed self-attention to learned features. On the self-attention map, we observe strong activation at the centers and edges of all nuclei, including unlabeled nuclei. On this basis, our region-proposal module propagates partial annotations to the whole image and proposes effective bounding boxes for the bounding box-regression and binary mask-generation modules. Our method effectively learns from unlabeled regions thereby improving detection performance. We test our method with various nuclear images. When trained with only 1/4 of the nuclei annotated, our approach retains a detection accuracy comparable to that from training with fully annotated data. Moreover, our method can serve as a bootstrapping step to create full annotations of datasets, iteratively generating and correcting annotations until a predetermined coverage and accuracy are reached. The source code is available at https://github.com/feng-lab/nuclei. INDEX TERMS Nucleus segmentation, deep learning, convolutional neural networks, few-exemplar learning, semisupervised learning, computer-assisted annotating.

show abstract

“…3. Furthermore, all of our models, even those trained with fewer annotations, show better performances in terms of the NCS than do those (approximately 0.6) reported in [31].…”

Section: B Overall Performancementioning

confidence: 63%

“…We evaluated the proposed method on the nucleus dataset published in [31]. The dataset contains fully annotated DAPI-stained cell nuclei, and we refer to it herein as DH100.…”

Section: Resultsmentioning

confidence: 99%

“…Normalized coverage score (NCS) values were used in [31] to evaluate the matching quality between the ground truth and the predicted segmentation. The NCS is defined as…”

Section: A Evaluationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Robust Nucleus Detection With Partially Labeled Exemplars

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Cropped images of the dorsal striatum (962 × 1236 pixels) were analyzed in two ways. The number of Fos transcripts per image was analyzed by counting the number of fluorescent dots in the Cy5 channel using SpotsInNucleiBot developed in Matlab (Cicconet et al, 2017). The same sigma and threshold values were used for all samples.…”

Section: Methodsmentioning

confidence: 99%

Reduction of repetitive behavior by co-administration of adenosine receptor agonists in C58 mice

Lewis

Rajpal

Muehlmann

2019

Pharmacology Biochemistry and Behavior

View full text Add to dashboard Cite

Repetitive behaviors are diagnostic for autism spectrum disorder (ASD) and commonly observed in other neurodevelopmental disorders. Currently, there are no effective pharmacological treatments for repetitive behavior in these clinical conditions. This is due to the lack of information about the specific neural circuitry that mediates the development and expression of repetitive behavior. Our previous work in mouse models has linked repetitive behavior to decreased activation of the subthalamic nucleus, a brain region in the indirect and hyperdirect pathways in the basal ganglia circuitry. The present experiments were designed to further test our hypothesis that pharmacological activation of the indirect pathway would reduce repetitive behavior. We used a combination of adenosine A1 and A2A receptor agonists that have been shown to alter the firing frequency of dorsal striatal neurons within the indirect pathway of the basal ganglia. This drug combination markedly and selectively reduced repetitive behavior in both male and female C58 mice over a six-hour period, an effect that required both A1 and A2A agonists as neither alone reduced repetitive behavior. The adenosine A1 and A2A receptor agonist combination also significantly increased the number of Fos transcripts and Fos positive cells in dorsal striatum. Fos induction was found in both direct and indirect pathway neurons suggesting that the drug combination restored the balance of activation across these complementary basal ganglia pathways. The adenosine A1 and A2A receptor agonist combination also maintained its effectiveness in reducing repetitive behavior over a 7-day period. These findings point to novel potential therapeutic targets for development of drug therapies for repetitive behavior in clinical disorders.

show abstract