Benchmarking of deep learning algorithms for 3D instance segmentation of confocal image datasets.

Segmenting three-dimensional (3D) microscopy images is essential for understanding phenomena like morphogenesis, cell division, cellular growth, and genetic expression patterns. Recently, deep learning (DL) pipelines have been developed, which claim to provide high accuracy segmentation of cellular images and are increasingly considered as the state of the art for image segmentation problems. However, it remains difficult to define their relative performances as the concurrent diversity and lack of uniform evaluation strategies makes it difficult to know how their results compare. In this paper, we first made an inventory of the available DL methods for 3D cell segmentation. We next implemented and quantitatively compared a number of representative DL pipelines, alongside a highly efficient non-DL method named MARS. The DL methods were trained on a common dataset of 3D cellular confocal microscopy images. Their segmentation accuracies were also tested in the presence of different image artifacts. A specific method for segmentation quality evaluation was adopted, which isolates segmentation errors due to under- or oversegmentation. This is complemented with a 3D visualization strategy for interactive exploration of segmentation quality. Our analysis shows that the DL pipelines have different levels of accuracy. Two of them, which are end-to-end 3D and were originally designed for cell boundary detection, show high performance and offer clear advantages in terms of adaptability to new data.

A multiscale analysis of early flower development in Arabidopsis provides an integrated view of molecular regulation and growth control.

We have analyzed the link between the gene regulation and growth during the early stages of flower development in Arabidopsis. Starting from time-lapse images, we generated a 4D atlas of early flower development, including cell lineage, cellular growth rates, and the expression patterns of regulatory genes. This information was introduced in MorphoNet, a web-based platform. Using computational models, we found that the literature-based molecular network only explained a minority of the gene expression patterns. This was substantially improved by adding regulatory hypotheses for individual genes. Correlating growth with the combinatorial expression of multiple regulators led to a set of hypotheses for the action of individual genes in morphogenesis. This identified the central factor LEAFY as a potential regulator of heterogeneous growth, which was supported by quantifying growth patterns in a leafy mutant. By providing an integrated view, this atlas should represent a fundamental step toward mechanistic models of flower development.

Tissue folding at the organ-meristem boundary results in nuclear compression and chromatin compaction.

Artificial mechanical perturbations affect chromatin in animal cells in culture. Whether this is also relevant to growing tissues in living organisms remains debated. In plants, aerial organ emergence occurs through localized outgrowth at the periphery of the shoot apical meristem, which also contains a stem cell niche. Interestingly, organ outgrowth has been proposed to generate compression in the saddle-shaped organ-meristem boundary domain. Yet whether such growth-induced mechanical stress affects chromatin in plant tissues is unknown. Here, by imaging the nuclear envelope in vivo over time and quantifying nucleus deformation, we demonstrate the presence of active nuclear compression in that domain. We developed a quantitative pipeline amenable to identifying a subset of very deformed nuclei deep in the boundary and in which nuclei become gradually narrower and more elongated as the cell contracts transversely. In this domain, we find that the number of chromocenters is reduced, as shown by chromatin staining and labeling, and that the expression of linker histone H1.3 is induced. As further evidence of the role of forces on chromatin changes, artificial compression with a MicroVice could induce the ectopic expression of H1.3 in the rest of the meristem. Furthermore, while the methylation status of chromatin was correlated with nucleus deformation at the meristem boundary, such correlation was lost in the h1.3 mutant. Altogether, we reveal that organogenesis in plants generates compression that is able to have global effects on chromatin in individual cells.

Multimodal characterization of acid-pretreated poplar reveals spectral and structural parameters strongly correlate with saccharification.

Lignocellulose biomass can be transformed into sustainable chemicals, materials and energy but its natural recalcitrance requires the use of pretreatment to enhance subsequent catalytic steps. Dilute acid pretreatment is one of the most common and efficient ones, however its impact has not yet been investigated simultaneously at nano- and cellular-scales. Poplar samples have been pretreated by dilute acid at different controlled severities, then characterized by combined structural and spectral techniques (scanning electron microscopy, confocal microscopy, autofluorescence, fluorescence lifetime, Raman). Results show that pretreatment favours lignin depolymerization until severity of 2.4-2.5 while at severity of 2.7 lignin seems to repolymerize as revealed by broadening of autofluorescence spectrum and strong decrease in fluorescence lifetime. Importantly, both nano-scale and cellular-scale markers can predict hydrolysis yield of pretreated samples, highlighting some connections in the multiscale recalcitrance of lignocellulose.

Phyllotactic regularity requires the Paf1 complex in Arabidopsis.

In plants, aerial organs are initiated at stereotyped intervals, both spatially (every 137° in a pattern called phyllotaxis) and temporally (at prescribed time intervals called plastochrons). To investigate the molecular basis of such regularity, mutants with altered architecture have been isolated. However, most of them only exhibit plastochron defects and/or produce a new, albeit equally reproducible, phyllotactic pattern. This leaves open the question of a molecular control of phyllotaxis regularity. Here, we show that phyllotaxis regularity depends on the function of VIP proteins, components of the RNA polymerase II-associated factor 1 complex (Paf1c). Divergence angles between successive organs along the stem exhibited increased variance in vip3-1 and vip3-2 compared with the wild type, in two different growth conditions. Similar results were obtained with the weak vip3-6 allele and in vip6, a mutant for another Paf1c subunit. Mathematical analysis confirmed that these defects could not be explained solely by plastochron defects. Instead, increased variance in phyllotaxis in vip3 was observed at the meristem and related to defects in spatial patterns of auxin activity. Thus, the regularity of spatial, auxin-dependent, patterning at the meristem requires Paf1c.

Fluctuations of the transcription factor ATML1 generate the pattern of giant cells in the Arabidopsis sepal.

Multicellular development produces patterns of specialized cell types. Yet, it is often unclear how individual cells within a field of identical cells initiate the patterning process. Using live imaging, quantitative image analyses and modeling, we show that during Arabidopsis thaliana sepal development, fluctuations in the concentration of the transcription factor ATML1 pattern a field of identical epidermal cells to differentiate into giant cells interspersed between smaller cells. We find that ATML1 is expressed in all epidermal cells. However, its level fluctuates in each of these cells. If ATML1 levels surpass a threshold during the G2 phase of the cell cycle, the cell will likely enter a state of endoreduplication and become giant. Otherwise, the cell divides. Our results demonstrate a fluctuation-driven patterning mechanism for how cell fate decisions can be initiated through a random yet tightly regulated process.

Cell size and growth regulation in the Arabidopsis thaliana apical stem cell niche.

Cell size and growth kinetics are fundamental cellular properties with important physiological implications. Classical studies on yeast, and recently on bacteria, have identified rules for cell size regulation in single cells, but in the more complex environment of multicellular tissues, data have been lacking. In this study, to characterize cell size and growth regulation in a multicellular context, we developed a 4D imaging pipeline and applied it to track and quantify epidermal cells over 3-4 d in Arabidopsis thaliana shoot apical meristems. We found that a cell size checkpoint is not the trigger for G2/M or cytokinesis, refuting the unexamined assumption that meristematic cells trigger cell cycle phases upon reaching a critical size. Our data also rule out models in which cells undergo G2/M at a fixed time after birth, or by adding a critical size increment between G2/M transitions. Rather, cell size regulation was intermediate between the critical size and critical increment paradigms, meaning that cell size fluctuations decay by ∼75% in one generation compared with 100% (critical size) and 50% (critical increment). Notably, this behavior was independent of local cell-cell contact topologies and of position within the tissue. Cells grew exponentially throughout the first >80% of the cell cycle, but following an asymmetrical division, the small daughter grew at a faster exponential rate than the large daughter, an observation that potentially challenges present models of growth regulation. These growth and division behaviors place strong constraints on quantitative mechanistic descriptions of the cell cycle and growth control.

A stochastic multicellular model identifies biological watermarks from disorders in self-organized patterns of phyllotaxis.

Exploration of developmental mechanisms classically relies on analysis of pattern regularities. Whether disorders induced by biological noise may carry information on building principles of developmental systems is an important debated question. Here, we addressed theoretically this question using phyllotaxis, the geometric arrangement of plant aerial organs, as a model system. Phyllotaxis arises from reiterative organogenesis driven by lateral inhibitions at the shoot apex. Motivated by recurrent observations of disorders in phyllotaxis patterns, we revisited in depth the classical deterministic view of phyllotaxis. We developed a stochastic model of primordia initiation at the shoot apex, integrating locality and stochasticity in the patterning system. This stochastic model recapitulates phyllotactic patterns, both regular and irregular, and makes quantitative predictions on the nature of disorders arising from noise. We further show that disorders in phyllotaxis instruct us on the parameters governing phyllotaxis dynamics, thus that disorders can reveal biological watermarks of developmental systems.

An epidermis-driven mechanism positions and scales stem cell niches in plants.

How molecular patterning scales to organ size is highly debated in developmental biology. We explore this question for the characteristic gene expression domains of the plant stem cell niche residing in the shoot apical meristem. We show that a combination of signals originating from the epidermal cell layer can correctly pattern the key gene expression domains and notably leads to adaptive scaling of these domains to the size of the tissue. Using live imaging, we experimentally confirm this prediction. The identified mechanism is also sufficient to explain de novo stem cell niches in emerging flowers. Our findings suggest that the deformation of the tissue transposes meristem geometry into an instructive scaling and positional input for the apical plant stem cell niche.

Meristem size contributes to the robustness of phyllotaxis in Arabidopsis.

Using the plant model Arabidopsis, the relationship between day length, the size of the shoot apical meristem, and the robustness of phyllotactic patterns were analysed. First, it was found that reducing day length leads to an increased meristem size and an increased number of alterations in the final positions of organs along the stem. Most of the phyllotactic defects could be related to an altered tempo of organ emergence, while not affecting the spatial positions of organ initiations at the meristem. A correlation was also found between meristem size and the robustness of phyllotaxis in two accessions (Col-0 and WS-4) and a mutant (clasp-1), independent of growth conditions. A reduced meristem size in clasp-1 was even associated with an increased robustness of the phyllotactic pattern, beyond what is observed in the wild type. Interestingly it was also possible to modulate the robustness of phyllotaxis in these different genotypes by changing day length. To conclude, it is shown first that robustness of the phyllotactic pattern is not maximal in the wild type, suggesting that, beyond its apparent stereotypical order, the robustness of phyllotaxis is regulated. Secondly, a role for day length in the robustness of the phyllotaxis was also identified, thus providing a new example of a link between patterning and environment in plants. Thirdly, the experimental results validate previous model predictions suggesting a contribution of meristem size in the robustness of phyllotaxis via the coupling between the temporal sequence and spatial pattern of organ initiations.

Cytokinin signalling inhibitory fields provide robustness to phyllotaxis.

How biological systems generate reproducible patterns with high precision is a central question in science. The shoot apical meristem (SAM), a specialized tissue producing plant aerial organs, is a developmental system of choice to address this question. Organs are periodically initiated at the SAM at specific spatial positions and this spatiotemporal pattern defines phyllotaxis. Accumulation of the plant hormone auxin triggers organ initiation, whereas auxin depletion around organs generates inhibitory fields that are thought to be sufficient to maintain these patterns and their dynamics. Here we show that another type of hormone-based inhibitory fields, generated directly downstream of auxin by intercellular movement of the cytokinin signalling inhibitor ARABIDOPSIS HISTIDINE PHOSPHOTRANSFER PROTEIN 6 (AHP6), is involved in regulating phyllotactic patterns. We demonstrate that AHP6-based fields establish patterns of cytokinin signalling in the meristem that contribute to the robustness of phyllotaxis by imposing a temporal sequence on organ initiation. Our findings indicate that not one but two distinct hormone-based fields may be required for achieving temporal precision during formation of reiterative structures at the SAM, thus indicating an original mechanism for providing robustness to a dynamic developmental system.

Pattern identification and characterization reveal permutations of organs as a key genetically controlled property of post-meristematic phyllotaxis.

In vascular plants, the arrangement of organs around the stem generates geometric patterns called phyllotaxis. In the model plant, Arabidopsis thaliana, as in the majority of species, single organs are initiated successively at a divergence angle from the previous organ close to the canonical angle of 137.5°, producing a Fibonacci spiral. Given that little is known about the robustness of these geometric arrangements, we undertook to characterize phyllotaxis by measuring divergence angles between organs along the stems of wild-type and specific mutant plants with obvious defects in phyllotaxis. Sequences of measured divergence angles exhibit segments of non-canonical angles in both genotypes, albeit to a far greater extent in the mutant. We thus designed a pipeline of methods for analyzing these perturbations. The latent structure models used in this pipeline combine a non-observable model representing perturbation patterns (either a variable-order Markov chain or a combinatorial model) with von Mises distributions representing divergence angle uncertainty. We show that the segments of non-canonical angles in both wild-type and mutant plants can be explained by permutations in the order of insertion along the stem of two or three consecutive organs. The number of successive organs between two permutations reveals specific patterns that depend on the nature of the preceding permutation (2- or 3-permutation). We also highlight significant individual deviations from 137.5° in the level of baseline segments and a marked relationship between permutation of organs and defects in the elongation of the internodes between these organs. These results demonstrate that permutations are an intrinsic property of spiral phyllotaxis and that their occurrence is genetically regulated.