The intra-annual dynamics of wood formation, which involves the passage of newly produced cells through three successive differentiation phases (division, enlargement, and wall thickening) to reach the final functional mature state, has traditionally been described in conifers as three delayed bell-shaped curves followed by an S-shaped curve. Here the classical view represented by the ‘Gompertz function (GF) approach’ was challenged using two novel approaches based on parametric generalized linear models (GLMs) and ‘data-driven’ generalized additive models (GAMs). These three approaches (GFs, GLMs, and GAMs) were used to describe seasonal changes in cell numbers in each of the xylem differentiation phases and to calculate the timing of cell development in three conifer species [Picea abies (L.), Pinus sylvestris L., and Abies alba Mill.]. GAMs outperformed GFs and GLMs in describing intra-annual wood formation dynamics, showing two left-skewed bell-shaped curves for division and enlargement, and a right-skewed bimodal curve for thickening. Cell residence times progressively decreased through the season for enlargement, whilst increasing late but rapidly for thickening. These patterns match changes in cell anatomical features within a tree ring, which allows the separation of earlywood and latewood into two distinct cell populations. A novel statistical approach is presented which renews our understanding of xylogenesis, a dynamic biological process in which the rate of cell production interplays with cell residence times in each developmental phase to create complex seasonal patterns.

The emergence of the characteristic tree-ring pattern during xylogenesis is commonly thought to be controlled by a gradient of morphogen (auxin, TDIF peptide...). We show that this hypothesis accounts for several developmental aspects of wood formation, but not for the final anatomical structure

Quantitative plant biology is an interdisciplinary field that builds on a long history of biomathematics and biophysics. Today, thanks to high spatiotemporal resolution tools and computational modelling, it sets a new standard in plant science. Acquired data, whether molecular, geometric or mechanical, are quantified, statistically assessed and integrated at multiple scales and across fields. They feed testable predictions that, in turn, guide further experimental tests. Quantitative features such as variability, noise, robustness, delays or feedback loops are included to account for the inner dynamics of plants and their interactions with the environment. Here, we present the main features of this ongoing revolution, through new questions around signalling networks, tissue topology, shape plasticity, biomechanics, bioenergetics, ecology and engineering. In the end, quantitative plant biology allows us to question and better understand our interactions with plants. In turn, this field opens the door to transdisciplinary projects with the society, notably through citizen science.

In conifers, xylogenesis produces during a growing season a very characteristic tree-ring structure: large thin-walled earlywood cells followed by narrow thick-walled latewood cells. Although many factors influence the dynamics of differentiation and the final dimensions of xylem cells, the associated patterns of variation remain very stable from one year to the next. While radial growth is characterised by an S-shaped curve, the widths of xylem differentiation zones exhibit characteristic skewed bell-shaped curves. These elements suggest a strong internal control of xylogenesis. It has long been hypothesised that much of this regulation relies on a morphogenetic gradient of auxin. However, recent modelling works have shown that while this hypothesis could account for the dynamics of stem radial growth and the zonation of the developing xylem, it failed to reproduce the characteristic tree-ring structure. Here we investigated the hypothesis of a regulation by a crosstalk between auxin and a second biochemical signal, using computational morphodynamics. We found that, in conifers, such a crosstalk is sufficient to simulate the characteristic features of wood formation dynamics, as well as the resulting tree-ring structure. In this model, auxin controls cell enlargement rates while another signal (e.g., cytokinin, TDIF) drives cell division and auxin polar transport.