The conversion of lignocellulosic biomass from plant cell walls into bioproducts can contribute to reducing dependence on fossil sources and achieving sustainable development. Biotechnological conversion of lignocellulosic biomass has several advantages over other conversion approaches such as thermochemical and chemical conversions. These advantages include improved efficiency and specificity for desired products, ecological compatibility and reduced toxicity. Enzymatic transformation is a key step in biotechnological conversion. To achieve a cost-effective conversion, a comprehensive understanding of cell wall enzymatic hydrolysis is required. Despite progress, the enzymatic hydrolysis at microscale is comparatively understudied and lacks comprehensive investigation. Addressing this gap requires collection of time-lapse image datasets of cell wall enzymatic hydrolysis which is a technically demanding task. Furthermore, accurate processing of the time-lapse images to identify and track individual cell walls is particularly challenging, notably because of the sample drift present in the images. Recently, an efficient image processing pipeline, called AIMTrack, has been developed which uses an enhanced divide-and-conquer strategy to divide time-lapse images into clusters whose sizes are dynamically adjusted to the deconstruction extent. The image registrations are then limited to clusters and the resulting transformations are combined to correct sample drift across time-lapse images. Subsequently AIMTrack provides segmentation of time-lapse images where voxels belonging to the same cell walls are labelled with a unique identifier. The time-lapse image datasets presented here consist of time-lapse images of spruce wood cell walls acquired during enzymatic hydrolysis using a cellulolytic enzyme cocktail at two enzyme loadings of 15 and 30 FPU/g biomass. Control time-lapse datasets which are acquired under the identical conditions, but without addition of enzymes, are also included. Both control and hydrolysis datasets are processed using AIMTrack to track the cell walls from time-lapse images. The generated segmentations are also provided. (c) 2026 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license

Enzymatic hydrolysis of plant cell walls into fermentable sugars is a critical step in biotechnological conversion, yet efficiency is limited by cell wall recalcitrance. Predicting conversion yields of cell wall-derived sugars during hydrolysis is challenging due to the complex underlying mechanisms and the labor-intensive nature of conventional methods. This study introduces an innovative pipeline that accurately quantifies cell wall autofluorescence intensity and morphological descriptors during enzymatic hydrolysis. The pipeline incorporates a novel adaptive drift compensation strategy that dynamically adjusts to the progression and extent of deconstruction, ensuring robust analysis. Applied to time-lapse images of spruce wood enzymatic deconstruction, the pipeline revealed strong negative correlations between conversion yields during hydrolysis and both the dynamics of cell wall autofluorescence intensity and morphological descriptors. Phase-specific analysis uncovered distinct correlation patterns dependent on hydrolysis stage and sugar type. This nondestructive pipeline eliminates the need for extensive sampling and time-consuming chemical assays, establishing plant cell wall autofluorescence and morphological descriptors as accurate predictive real-time biomarkers of dynamics of sugar conversion yields. The findings provide a framework for accelerating the development of optimized biotechnological conversion processes.

Achieving an economically viable transformation of plant cell walls into bioproducts requires a comprehensive understanding of enzymatic deconstruction. Microscale quantitative analysis offers a relevant approach to enhance our understanding of cell wall hydrolysis, but becomes challenging under high deconstruction conditions. This study comprehensively addresses the challenges of quantifying the impact of extensive enzymatic deconstruction on plant cell wall at microscale. Investigation of highly deconstructed spruce wood provided spatial profiles of cell walls during hydrolysis with remarkable precision. A distinct cell wall autofluorescence distribution pattern marking enzymatic hydrolysis along with an asynchronous impact of hydrolysis on cell wall structure, with cell wall volume reduction preceding cell wall accessible surface area decrease, were revealed. This study provides novel insights into enzymatic deconstruction of cell wall at under-investigated cell scale, and a robust computational pipeline applicable to diverse biomass species and pretreatment types for assessing hydrolysis impact and efficiency.

Understanding lignocellulosic biomass resistance to enzymatic deconstruction is crucial for its sustainable conversion into bioproducts. Despite scientific advances, quantitative morphological analysis of plant deconstruction at cell and tissue scales remains under-explored. In this study, an original pipeline is devised, involving four-dimensional (space  + time) fluorescence confocal imaging, and a novel computational tool, to track and quantify deconstruction at cell and tissue scales. By applying this pipeline to poplar wood, dynamics of cellular parameters was computed and cellulose conversion during enzymatic deconstruction was measured. Results showed that enzymatic deconstruction predominantly impacts cell wall volume rather than surface area. Additionally, a negative correlation was observed between pre-hydrolysis compactness measures and volumetric cell wall deconstruction rate, whose strength was modulated by enzymatic activity. Results also revealed a strong positive correlation between average volumetric cell wall deconstruction rate and cellulose conversion rate. These findings link key deconstruction parameters across nano and micro scales.

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.