Fast and Efficient Root Phenotyping via Pose Estimation.

Image segmentation is commonly used to estimate the location and shape of plants and their external structures. Segmentation masks are then used to localize landmarks of interest and compute other geometric features that correspond to the plant's phenotype. Despite its prevalence, segmentation-based approaches are laborious (requiring extensive annotation to train) and error-prone (derived geometric features are sensitive to instance mask integrity). Here, we present a segmentation-free approach that leverages deep learning-based landmark detection and grouping, also known as pose estimation. We use a tool originally developed for animal motion capture called SLEAP (Social LEAP Estimates Animal Poses) to automate the detection of distinct morphological landmarks on plant roots. Using a gel cylinder imaging system across multiple species, we show that our approach can reliably and efficiently recover root system topology at high accuracy, few annotated samples, and faster speed than segmentation-based approaches. In order to make use of this landmark-based representation for root phenotyping, we developed a Python library (sleap-roots) for trait extraction directly comparable to existing segmentation-based analysis software. We show that pose-derived root traits are highly accurate and can be used for common downstream tasks including genotype classification and unsupervised trait mapping. Altogether, this work establishes the validity and advantages of pose estimation-based plant phenotyping. To facilitate adoption of this easy-to-use tool and to encourage further development, we make sleap-roots, all training data, models, and trait extraction code available at: https://github.com/talmolab/sleap-roots and https://osf.io/k7j9g/.

Fast and efficient root phenotyping via pose estimation.

Image segmentation is commonly used to estimate the location and shape of plants and their external structures. Segmentation masks are then used to localize landmarks of interest and compute other geometric features that correspond to the plant's phenotype. Despite its prevalence, segmentation-based approaches are laborious (requiring extensive annotation to train), and error-prone (derived geometric features are sensitive to instance mask integrity). Here we present a segmentation-free approach which leverages deep learning-based landmark detection and grouping, also known as pose estimation. We use a tool originally developed for animal motion capture called SLEAP (Social LEAP Estimates Animal Poses) to automate the detection of distinct morphological landmarks on plant roots. Using a gel cylinder imaging system across multiple species, we show that our approach can reliably and efficiently recover root system topology at high accuracy, few annotated samples, and faster speed than segmentation-based approaches. In order to make use of this landmark-based representation for root phenotyping, we developed a Python library (sleap-roots) for trait extraction directly comparable to existing segmentation-based analysis software. We show that landmark-derived root traits are highly accurate and can be used for common downstream tasks including genotype classification and unsupervised trait mapping. Altogether, this work establishes the validity and advantages of pose estimation-based plant phenotyping. To facilitate adoption of this easy-to-use tool and to encourage further development, we make sleap-roots, all training data, models, and trait extraction code available at: https://github.com/talmolab/sleap-roots and https://osf.io/k7j9g/.

Foodborne agents associated with the consumption of aquaculture catfish.

In the Food, Conservation, and Energy Act (Farm Bill) of 2008, Congress amended the Federal Meat Inspection Act to provide that catfish be inspected by the U.S. Department of Agriculture's Food Safety and Inspection Service (FSIS). As part of the development of its inspection program, the FSIS conducted an assessment of the food safety risk associated with consuming farm-raised catfish. To thoroughly identify hazards for consideration in the risk assessment, the scientific literature was surveyed for all potential agents that have been linked to illness associated with farm-raised catfish consumption. A review of microbial hazards suggested that Salmonella is the foodborne pathogen most likely to be associated with catfish, but the impact of other pathogens remains unclear. This review also summarizes the current data available on chemical residues in catfish, including pesticides and heavy metals, and any regulatory levels that have been established for these compounds. The current usage of veterinary drugs in aquaculture also is outlined, including information on unapproved usage of drugs in catfish.

Characterization of the Six1 homeobox gene in normal mammary gland morphogenesis.

BACKGROUND: The Six1 homeobox gene is highly expressed in the embryonic mammary gland, continues to be expressed in early postnatal mammary development, but is lost when the mammary gland differentiates during pregnancy. However, Six1 is re-expressed in breast cancers, suggesting that its re-instatement in the adult mammary gland may contribute to breast tumorigenesis via initiating a developmental process out of context. Indeed, recent studies demonstrate that Six1 overexpression in the adult mouse mammary gland is sufficient for initiating invasive carcinomas, and that its overexpression in xenograft models of mammary cancer leads to metastasis. These data demonstrate that Six1 is causally involved in both breast tumorigenesis and metastasis, thus raising the possibility that it may be a viable therapeutic target. However, because Six1 is highly expressed in the developing mammary gland, and because it has been implicated in the expansion of mammary stem cells, targeting Six1 as an anti-cancer therapy may have unwanted side effects in the breast. RESULTS: We sought to determine the role of Six1 in mammary development using two independent mouse models. To study the effect of Six1 loss in early mammary development when Six1 is normally expressed, Six1-/- embryonic mammary glands were transplanted into Rag1-/- mice. In addition, to determine whether Six1 downregulation is required during later stages of development to allow for proper differentiation, we overexpressed Six1 during adulthood using an inducible, mammary-specific transgenic mouse model. Morphogenesis of the mammary gland occurred normally in animals transplanted with Six1-/- embryonic mammary glands, likely through the redundant functions of other Six family members such as Six2 and Six4, whose expression was increased in response to Six1 loss. Surprisingly, inappropriate expression of Six1 in the adult mammary gland, when levels are normally low to absent, did not inhibit normal mammary differentiation during pregnancy or lactation. CONCLUSIONS: Six1 is not critical for normal mammary gland development, since neither loss nor inappropriate overexpression of Six1 adversely affects normal mammary gland development or function. However, as both Six2 and Six4 levels are increased in Six1-/- mammary glands, we postulate that these Six family members are functionally redundant in the gland, as is true of many homeobox gene families. This data, in conjunction with recent findings that Six1 is capable of promoting breast cancer initiation and progression, suggest that Six1 may serve as a reasonable chemotherapeutic target in a clinical setting, particularly for those women diagnosed with breast cancer in their childbearing years.

Six1 expands the mouse mammary epithelial stem/progenitor cell pool and induces mammary tumors that undergo epithelial-mesenchymal transition.

Six1 is a developmentally regulated homeoprotein with limited expression in most normal adult tissues and frequent misexpression in a variety of malignancies. Here we demonstrate, using a bitransgenic mouse model, that misexpression of human Six1 in adult mouse mammary gland epithelium induces tumors of multiple histological subtypes in a dose-dependent manner. The neoplastic lesions induced by Six1 had an in situ origin, showed diverse differentiation, and exhibited progression to aggressive malignant neoplasms, as is often observed in human carcinoma of the breast. Strikingly, the vast majority of Six1-induced tumors underwent an epithelial-mesenchymal transition (EMT) and expressed multiple targets of activated Wnt signaling, including cyclin D1. Interestingly, Six1 and cyclin D1 coexpression was found to frequently occur in human breast cancers and was strongly predictive of poor prognosis. We further show that Six1 promoted a stem/progenitor cell phenotype in the mouse mammary gland and in Six1-driven mammary tumors. Our data thus provide genetic evidence for a potent oncogenic role for Six1 in mammary epithelial neoplasia, including promotion of EMT and stem cell-like features.

The six family of homeobox genes in development and cancer.

The homeobox gene superfamily encodes transcription factors that act as master regulators of development through their ability to activate or repress a diverse range of downstream target genes. Numerous families exist within the homeobox gene superfamily, and are classified on the basis of conservation of their homeodomains as well as additional motifs that contribute to DNA binding and to interactions with other proteins. Members of one such family, the Six family, form a transcriptional complex with Eya and Dach proteins, and together these proteins make up part of the retinal determination network first identified in Drosophila. This network is highly conserved in both invertebrate and vertebrate species, where it influences the development of numerous organs in addition to the eye, primarily through regulation of cell proliferation, survival, migration, and invasion. Mutations in Six, Eya, and Dach genes have been identified in a variety of human genetic disorders, demonstrating their critical role in human development. In addition, aberrant expression of Six, Eya, and Dach occurs in numerous human tumors, and Six1, in particular, plays a causal role both in tumor initiation and in metastasis. Emerging evidence for the importance of Six family members and their cofactors in numerous human tumors suggests that targeting of this complex may be a novel and powerful means to inhibit both tumor growth and progression.

Nkx2.2 regulates cell fate choice in the enteroendocrine cell lineages of the intestine.

Nkx2.2 is a homeodomain-containing transcription factor essential for pancreatic islet cell specification. In this study we investigate the role of Nkx2.2 within the small intestine. We have determined that Nkx2.2 is expressed at the onset of intestinal epithelial cell differentiation in specific intestinal cell populations, including a subset of enteroendocrine cells. Similar to its role in the pancreatic islet, Nkx2.2 regulates cell fate choices within the intestinal enteroendocrine population; in the Nkx2.2 null mice, several hormone-producing enteroendocrine cell populations are absent or reduced and the ghrelin-producing cell population is upregulated. The remaining intestinal cell populations, including the paneth cells, goblet cells, and enterocytes appear to be unaffected by the loss of Nkx2.2. Furthermore, similar to the pancreatic islet, Nkx2.2 appears to function upstream of Pax6 in regulating intestinal cell fates; Pax6 mRNA and protein expression is decreased in the Nkx2.2 null mice. These studies identify a novel role for Nkx2.2 in intestinal endocrine cell development and reveal the regulatory similarities between cell type specification in the pancreatic islet and in the enteroendocrine population of the intestine.