Crop biomass offers crucial insights into plant health and yield, making it essential for crop science, farming systems, and agricultural research. However, current measurement methods, which are labor-intensive, destructive, and imprecise, hinder large-scale quantification of this trait. To address this limitation, we present a biomass prediction network (BioNet), designed for adaptation across different data modalities, including point clouds and drone imagery. Our BioNet, utilizing a sparse 3D convolutional neural network (CNN) and a transformer-based prediction module, processes point clouds and other 3D data representations to predict biomass. To further extend BioNet for drone imagery, we integrate a neural feature field (NeFF) module, enabling 3D structure reconstruction and the transformation of 2D semantic features from vision foundation models into the corresponding 3D surfaces. For the point cloud modality, BioNet demonstrates superior performance on two public datasets, with an approximate 6.1% relative improvement (RI) over the state-of-the-art. In the RGB image modality, the combination of BioNet and NeFF achieves a 7.9% RI. Additionally, the NeFF-based approach utilizes inexpensive, portable drone-mounted cameras, providing a scalable solution for large field applications.

Traditional methods utilising data-independent acquisition mass spectrometry (DIA-MS) data for predictions depend on database searches against predefined spectral libraries for characterisation and quantification of the proteomes, limiting scalability and adaptability across various applications. However, directly applying existing networks on DIA-MS data represented as images for end-to-end predictions struggles to mine a predictive pattern, due to non-uniform region importance across the image and divergences exhibited among different regions of the image. To overcome these limitations, we propose a new frame-work with two modules: i) a dynamic sampling module that identifies regions of interest from the DIA-MS image, constraining the network to focus on the most informative regions of the image only; ii) a mixture of experts module that sparsely routes the regions of interest to related expert networks, facilitating adaptive computation of region features. The region features are then fused for predictions. Experimentally, to benchmark our method, we collected a large DIA-MS dataset of wheat for flowering time prediction, and our approach significantly outperforms previous end-to-end methods, i. e., 0.171 R2 improvements.

This paper studies a convolutional masked image modeling approach for boosting downstream dense prediction tasks on pathology images. Our method is self-supervised, and entails two strategies in sequence. Considering features contained in the pathology images usually have a large spatial span, e.g., glands, we insert [MASK] tokens to the masked regions after the stem layer of the convolutional network for encoding unmasked pixels, which facilitates information propagation through masked regions for reconstructing unmasked pixels. Furthermore, the pathology images contain features that are represented in diverse affine shapes and color spaces. We, therefore, enforce the network to learn the affine and color invariant embedding by imposing transformation constraints between the unmasked image-encoded embedding and reconstruction targets. Our approach is simple but effective. With extensive experiments on standard benchmark datasets, we demonstrate superior transfer learning performance on downstream tasks over past state-of-the-art approaches.

Spatial transcriptomics (ST) is essential for understanding diseases and developing novel treatments. It measures the gene expression of each fine-grained area (i.e., different windows) in the tissue slide with low throughput. This paper proposes an exemplar guided graph network dubbed EGGN to accurately and efficiently predict gene expression from each window of a tissue slide image. We apply exemplar learning to dynamically boost gene expression prediction from nearest/similar exemplars of a given tissue slide image window. Our framework has three main components connected in a sequence: (i) an extractor to structure a feature space for exemplar retrievals; (ii) a graph construction strategy to connect windows and exemplars as a graph; (iii) a graph convolutional network backbone to process window and exemplar features, and a graph exemplar bridging block to adaptively revise the window features using its exemplars. Finally, we complete the gene expression prediction task with a simple attention-based prediction block. Experiments on standard benchmark datasets indicate the superiority of our approach when compared with past state-of-the-art methods. We release our code at https://github.com/Yan98/EGN.

Spatial transcriptomics (ST) is essential for understanding diseases and developing novel treatments. It measures gene expression of each fine-grained area (i.e., different windows) in the tissue slide with low throughput. This paper proposes an Exemplar Guided Network (EGN) to accurately and efficiently predict gene expression directly from each window of a tissue slide image. We apply exemplar learning to dynamically boost gene expression prediction from nearest/similar exemplars of a given tissue slide image window. Our EGN framework composes of three main components: 1) an extractor to structure a representation space for unsupervised exemplar retrievals; 2) a vision transformer (ViT) backbone to progressively extract representations of the input window; and 3) an Exemplar Bridging (EB) block to adaptively revise the intermediate ViT representations by using the nearest exemplars. Finally, we complete the gene expression prediction task with a simple attention-based prediction block. Experiments on standard benchmark datasets indicate the superiority of our approach when comparing with the past state-of-the-art (SOTA) methods.

With population growth and a shrinking rural workforce, agricultural technologies have become increasingly important. Above-ground biomass (AGB) is a key trait relevant to breeding, agronomy and crop physiology field experiments. However, measuring the biomass of a cereal plot requires cutting, drying and weighing processes, which are laborious, expensive and destructive tasks. This paper proposes a non-destructive and high-throughput method to predict biomass from field samples based on Light Detection and Ranging (LiDAR). Unlike previous methods that are based on the density of a point cloud or plant height, our biomass prediction network (BioNet) additionally considers plant structure. Our BioNet contains three modules: 1) a completion module to predict missing points due to canopy occlusion; 2) a regularization module to regularize the neural representation of the whole plot; and 3) a projection module to learn the salient structures from a bird’s eye view of the point cloud. An attention-based fusion block is used to achieve final biomass predictions. In addition, the complete dataset, including hand-measured biomass and LiDAR data, is made available to the community. Experiments show that our BioNet achieves ≈ 33% improvement over current state-of-the-art methods.

BACKGROUND: As a consequence of the evolutionary process, data collected from related species tend to be similar. This similarity by descent can obscure subtler signals in the data such as the evidence of constraint on variation due to shared selective pressures. In comparative sequence analysis, for example, sequence similarity is often used to illuminate important regions of the genome, but if the comparison is between closely related species, then similarity is the rule rather than the interesting exception. Furthermore, and perhaps worse yet, the contribution of a divergent third species may be masked by the strong similarity between the other two. Here we propose a remedy that weighs the contribution of each species according to its phylogenetic placement. RESULTS: We first solve the problem of summarizing data related by phylogeny, and we explain why an average should operate on the entire evolutionary trajectory that relates the data. This perspective leads to a new approach in which we define the average in terms of the phylogeny, using the data and a stochastic model to obtain a probability on evolutionary trajectories. With the assumption that the data evolve according to a Brownian motion process on the tree, we show that our evolutionary average can be computed as convex combination of the species data. Thus, our approach, called the BranchManager, defines both an average and a novel taxon weighting scheme. We compare the BranchManager to two other methods, demonstrating why it exhibits desirable properties. In doing so, we devise a framework for comparison and introduce the concept of a representative point at which the average is situated. CONCLUSION: The BranchManager uses as its representative point the phylogenetic center of mass, a choice which has both intuitive and practical appeal. Because our average is intrinsic to both the dataset and to the phylogeny, we expect it and its corresponding weighting scheme to be useful in all sorts of studies where interspecies data need to be combined. Obvious applications include evolutionary studies of morphology, physiology or behaviour, but quantitative measures such as sequence hydrophobicity and gene expression level are amenable to our approach as well. Other areas of potential impact include motif discovery and vaccine design. A Java implementation of the BranchManager is available for download, as is a script written in the statistical language R.