Polyploidy can be a critical factor for explaining plant trait variation, niche diversification, or speciation. However, inferring ploidy from silica-dried or historical samples using chromosome counts or flow cytometry is not possible, and scaling up ploidy estimation to population-level fresh contemporary samples can be challenging as well. Thus, we present a new method for estimating ploidy levels directly from sequencing data using machine learning; the Polyploid Population Genomics Tool Kit (PPGTK). The machine-learning approach is advantageous as it relaxes the assumptions of previous probabilistic methods and provides per-sample probabilities, allowing investigators to evaluate uncertainty in their system of interest.. We demonstrate performance and accuracy of the method on simulated and empirical data. Simulations showed above 99% accuracy, even for low coverage data, as long reads were mappable to the reference genome. For empirical analyses, we used target enrichment data from blueberry wild relatives (Vaccinium sect. Cyanococcus) and whole-genome data from sweetpotato wild relatives (Ipomoea ser. Batatas). Ploidy was recovered with 99% accuracy across 70 Vaccinium individuals and 97% across 82 Ipomoea individuals. Analysis of many individuals is fast and requires only a multisample VCF, which is presumably generated for the research anyway, and some samples of known ploidy for training the classifier. The approach implemented in PPGTK is promising for collections-based research as well, enabling ploidy classification of historical specimens based on present-day observations. The method is implemented in a new Python package as a single command that can run on a conventional laptop.
Why it matches plant phenotyping methods植物の倍数性という状態をシーケンスデータから推定する機械学習手法を開発し、シミュレーションおよび実データで精度検証している。Pythonパッケージとして実装され、手法自体が中心である。
abstractwe present a new method for estimating ploidy levels directly from sequencing data using machine learning; the Polyploid Population Genomics Tool Kit (PPGTK).
Reproduction assets foundThe paper's ploidy-classification method is implemented in the authors' public Python package PPGTK, with a specific release (v0.1.0-alpha) used for the manuscript's analyses. The empirical VCF/metadata datasets are promised on Dryad only 'upon acceptance' and thus are not yet actionable.Code · public11
VCFs and metadata needed to reproduce Vaccinium sect. Cyanococcus and Ipomoea ser.
372
Batatas analyses with PPGTK will be made available via Dryad upon acceptance. PPGTK is
373
available on GitHub, and release v0.1.0-alpha was the version used for analyses in this
374
manuscript (https://github.com/tileylab/PPGTK/releases/tag/v0.1.0-alpha). PPGTK currently has
375
other functions for calculating population genetic summary statistics, but the classify-ploidy
376
function implements the machine-learning method described in the manuscript.
377
378
.
CC-BY 4.0 International license
is made available under a
preprint (which was not certified by peer review) is the auOpen asset ↗tileylab/PPGTK · v0.1.0-alphapdf-raw-page:11 lines:1-24Code / dataset availability confirmedOpenAlex · Crossref · checked 14 Sept 2026
Abstract Fruit size, shape, color, and percent fruit rot are important quality traits for breeding cranberry ( Vaccinium macrocarpon Ait.). Image analysis can be used to measure these traits, but affordable hardware for standardized image capture and integrated user‐friendly software pipelines are lacking. Additionally, no image‐based method exists to estimate percent fruit rot, an otherwise tediously and subjectively measured trait. We created the BerryBox, a simple and inexpensive lightbox, camera mount, and accompanying software pipeline to standardize the capture and analysis of postharvest fruit images. Trained deep neural network models were highly accurate for segmenting sound fruit (F1 score: 99.4%) and detecting rotten fruit (F1 score: 98.5%). We applied the BerryBox to images of cranberries harvested across 3 years from a 156‐clone breeding population. Narrow‐sense heritability estimates of image‐based fruit color, shape, size, and percent fruit rot ranged from 0.37 to 0.95. Random subsampling showed that 25–30 berries per genotype were sufficient to describe the variation in the full dataset. We demonstrated the utility of BerryBox traits in a small‐scale genetic linkage mapping analysis, detecting significant marker–trait associations that coincided with those of traditionally measured traits. The BerryBox software was able to accurately segment fruit from images of blueberries without model retraining, showing its applicability to other similarly shaped fruits. The software pipeline and BerryBox materials and assembly instructions are publicly available for others to adopt for low‐cost image‐based phenotyping.
Why it matches plant phenotyping methodsクランベリー等の果実形質と腐敗率を画像から抽出する低コスト撮像装置・ソフトウェアパイプラインを開発し、精度検証と他果実への適用性評価を行った、中心的な植物フェノタイピング手法研究である。
abstractWe created the BerryBox, a simple and inexpensive lightbox, camera mount, and accompanying software pipeline to standardize the capture and analysis of postharvest fruit images.
Reproduction assets foundThe paper explicitly states public availability of the annotated image datasets (USDA Ag Data Commons DOI), R analysis scripts, the BerryBox Python software package with pre-trained models, and the model training code, all with author-provided public URLs.Dataset · publics (LOD)
score at a particular marker exceeded that computed at the
α = 0.05 level under null models generated via 1000 random
permutations.
2.8 Data, software, and equipment
instruction availability
The image datasets, along with annotations, are publicly
available through the USDA National Agricultural Library
Ag Data Commons (https://doi.org/10.15482/USDA.ADC/29853332). All analyses in this study were performed in R (v.
4.5.0; R Core Team, 2025). Scripts to replicate the analyses,
along with a list of materials for recreating the Berry-
Box, are available from the GitHub repository https://github.com/neyhartj/BerryBox_FruitPhenotyping. Software for run-
ning the image capture and analysis Open asset ↗10.15482/USDA.ADC/29853332pdf-raw-page:8 lines:1-125Code · publicable through the USDA National Agricultural Library
Ag Data Commons (https://doi.org/10.15482/USDA.ADC/29853332). All analyses in this study were performed in R (v.
4.5.0; R Core Team, 2025). Scripts to replicate the analyses,
along with a list of materials for recreating the Berry-
Box, are available from the GitHub repository https://github.com/neyhartj/BerryBox_FruitPhenotyping. Software for run-
ning the image capture and analysis software pipeline is
available as a Python package from the GitHub reposi-
tory https://github.com/NeyhartLab/berryboxai. The package
includes pre-trained models for berry segmentation and fruit
rot detection, and the code is available from https://github.com/NOpen asset ↗github.com/neyhartj/BerryBox_FruitPhenotypingpdf-raw-page:8 lines:1-125Code · public). Scripts to replicate the analyses,
along with a list of materials for recreating the Berry-
Box, are available from the GitHub repository https://github.com/neyhartj/BerryBox_FruitPhenotyping. Software for run-
ning the image capture and analysis software pipeline is
available as a Python package from the GitHub reposi-
tory https://github.com/NeyhartLab/berryboxai. The package
includes pre-trained models for berry segmentation and fruit
rot detection, and the code is available from https://github.com/NeyhartLab/berryboxai_training_public for training a
custom model using high-performance computing resources
or the widely available Google Colab environment (Rippner
et al., 2022).
3 RESULTOpen asset ↗github.com/NeyhartLab/berryboxaipdf-raw-page:8 lines:1-125Code · publiceyhartj/BerryBox_FruitPhenotyping. Software for run-
ning the image capture and analysis software pipeline is
available as a Python package from the GitHub reposi-
tory https://github.com/NeyhartLab/berryboxai. The package
includes pre-trained models for berry segmentation and fruit
rot detection, and the code is available from https://github.com/NeyhartLab/berryboxai_training_public for training a
custom model using high-performance computing resources
or the widely available Google Colab environment (Rippner
et al., 2022).
3 RESULTS
3.1 Deep learning model training
The trained berry segmentation model achieved an overall
accuracy of 98.9% and an F1 score of 99.4%. The fruit rot
detection mOpen asset ↗github.com/NeyhartLab/berryboxai_training_publicpdf-raw-page:8 lines:1-125Code / dataset availability confirmedOpenAlex · Crossref · checked 6 Sept 2026
Time to maturity and yield are important traits for highbush blueberry (Vaccinium corymbosum) breeding. Proper determination of the time to maturity of blueberry varieties and breeding lines informs the harvest window, ensuring that the fruits are harvested at optimum maturity and quality. On the other hand, high-yielding crops bring in high profits per acre of planting. Harvesting and quantifying the yield for each blueberry breeding accession are labor-intensive and impractical. Instead, visual ratings as an estimation of yield are often used as a faster way to quantify the yield, which is categorical and subjective. In this study, we developed and shared a high-throughput phenotyping method using neural networks to predict blueberry time to maturity and to provide a proxy for yield, overcoming the labor constraints of obtaining high-frequency data. We aim to facilitate further research in computer vision and precision agriculture by publishing the labeled image dataset and the trained model. In this research, true-color images of blueberry bushes were collected, annotated, and used to train a deep neural network object detection model [You Only Look Once (YOLOv11)] to detect mature and immature berries. Different versions of YOLOv11 were used, including nano, small, and medium, which had similar performance, while the medium version had slightly higher metrics. The YOLOv11m model shows strong performance for the mature berry class, with a precision of 0.90 and an F1 score of 0.90. The precision and recall for detecting immature berries were 0.81 and 0.79. The model was tested on 10 blueberry bushes by hand harvesting and weighing blueberries. The results showed that the model detects approximately 25% of the berries on the bushes, and the correlation coefficients between model-detected and hand-harvested traits were 0.66, 0.86, and 0.72 for mature fruit count, immature fruit count, and mature ratio, respectively. The model applied to 91 blueberry advance selections and categorized them into groups with diverse levels of maturity and productivity using principal component analysis (PCA). These results inform the harvest window and yield of these breeding lines with precision and objectivity through berry classification and quantification. This model will be helpful for blueberry breeders, enabling more efficient selection, and for growers, helping them accurately estimate optimal harvest windows. This open-source tool can potentially enhance research capabilities and agricultural productivity.
Why it matches plant phenotyping methodsブルーベリーの成熟度・収量 proxy を画像とニューラルネットワークで推定する高スループット表現型計測法を開発・検証し、モデルとラベル付きデータセットを共有しているため、方法が研究の中心である。
abstractwe developed and shared a high-throughput phenotyping method using neural networks to predict blueberry time to maturity and to provide a proxy for yield
Reproduction assets foundThe paper publishes its labeled blueberry image dataset on Zenodo (record 14014858) and its trained YOLOv11-based blueberry fruit counting model/code on GitHub (jeromemaleski/blueberry), both directly supporting the paper's phenotyping measurements and analysis.Dataset · public32. Zhang, J. Blueberry Images and Labels for YOLO Model Training. Zenodo. 2024. Available online: https://zenodo.org/records/Open asset ↗zenodopdf-page:14 lines:1-36Code / dataset availability confirmedEurope PMC · bioRxiv · checked 15 Sept 2026
We present a sensitive method for non-contact detection of chloroplast movements in leaves and other photosynthetic tissues, based on changes in the magnitude of cross-polarized reflectance. We examined changes in bidirectional red light reflectance during irradiation with blue light, known to trigger chloroplast relocations. Experiments on the model plant Arabidopsis thaliana , wild-type, and several mutants with disrupted chloroplast movements showed that the chloroplast avoidance response, induced by high blue light, led to a substantial increase in diffuse reflectance of unpolarized red light. The effects of the accumulation response in low blue light were the opposite. The specular reflectance of the leaf was unaffected by the chloroplast positioning. To further improve the specificity of the detection, we examined the effects of chloroplast relocations on the leaf reflectance of a linearly polarized incident beam. The greatest relative change associated with chloroplast movements was observed when the planes of polarization of the incident and detected beams were perpendicular. Further experiments revealed that the chloroplast positioning affected the magnitude of depolarization of light by the leaf. We applied the developed approach to examine chloroplast relocations in four angiosperm species collected in the field. The method allowed us to detect the chloroplast avoidance response in the green stems of bilberry, a sample not amenable to transmittance-based detection. Despite the importance of chloroplast movements for the optimization of photosynthetic efficiency and biomass production, high throughput reflectance-based methods are not routinely used for their detection. This method opens the possibility of non-invasive, non-contact detection of chloroplast relocations in a manner insensitive to the orientation of the leaf.
Why it matches plant phenotyping methods葉のクロロプラスト移動という植物状態を、偏光反射によって非接触・非侵襲的に検出する手法を開発し、複数種で適用しているため、植物フェノタイピング手法が研究の中心である。
abstractWe present a sensitive method for non-contact detection of chloroplast movements in leaves and other photosynthetic tissues, based on changes in the magnitude of cross-polarized reflectance.
Reproduction assets foundThe paper's Data availability statement openly deposits the paper's own reflectance/transmittance phenotype recordings (Arabidopsis WT/mutants and wild plants) on FigShare, and provides authors' public code: BeamJ (Java control software for the phenotyping setup) and openRayTracer (Mathematica ray-tracing package used,Dataset · publiced on the manuscript.
Conflict of interest
The authors declare no conflict of interest.
Funding
This study was supported by the National Science Centre Poland within the MINIATURA 4
project to P.H., number 2020/04/X/NZ4/01256.
Data availability
The data that support the findings of this study are openly available in FigShare at
https://doi.org/10.6084/m9.figshare.21082654 (reflectance and transmittance recordings for
Arabidopsis WT and mutants) and https://doi.org/10.6084/m9.figshare.24424843 (wild
plants). Java source code for the software is publicly available via GitHub at
https://github.com/pawelHerm/beamJ/tree/master/BeamJ. Wolfram Mathematica package for
ray tracing is available at httOpen asset ↗FigShare · 10.6084/m9.figshare.21082654pdf-raw-page:14 lines:1-47Dataset · publicthe National Science Centre Poland within the MINIATURA 4
project to P.H., number 2020/04/X/NZ4/01256.
Data availability
The data that support the findings of this study are openly available in FigShare at
https://doi.org/10.6084/m9.figshare.21082654 (reflectance and transmittance recordings for
Arabidopsis WT and mutants) and https://doi.org/10.6084/m9.figshare.24424843 (wild
plants). Java source code for the software is publicly available via GitHub at
https://github.com/pawelHerm/beamJ/tree/master/BeamJ. Wolfram Mathematica package for
ray tracing is available at https://github.com/plantPhotobiologyLab/openRayTracer.References
Banaś, A. K., Aggarwal, C., Łabuz, J., Sztatelman, O., Gabryś,Open asset ↗FigShare · 10.6084/m9.figshare.24424843pdf-raw-page:14 lines:1-47Code · publicthat support the findings of this study are openly available in FigShare at
https://doi.org/10.6084/m9.figshare.21082654 (reflectance and transmittance recordings for
Arabidopsis WT and mutants) and https://doi.org/10.6084/m9.figshare.24424843 (wild
plants). Java source code for the software is publicly available via GitHub at
https://github.com/pawelHerm/beamJ/tree/master/BeamJ. Wolfram Mathematica package for
ray tracing is available at https://github.com/plantPhotobiologyLab/openRayTracer.References
Banaś, A. K., Aggarwal, C., Łabuz, J., Sztatelman, O., Gabryś, H. (2012). Blue light
signalling in chloroplast movements. Journal of Experimental Botany, 63(4), 1559–
1574.
Baránková, B., LazáOpen asset ↗GitHub · pawelHerm/beamJpdf-raw-page:14 lines:1-47Code · publicuorescence. The filtered light was focused on a
photodetector (amplified silicon photodiode, PDA100A2, Thorlabs) with a plano-convex lens
(LA1074-A, Thorlabs). The angular size of the clear aperture of the collecting lens with
respect to the sample center was 0.019 steradian (calculated using our ray-tracing
Mathematica package https://github.com/plantPhotobiologyLab/openRayTracer). To control
the observation angle, the detector was mounted at the RBB300A/M rotation board
(Thorlabs). The experiments were performed with two angular positions of the polarizer: its
transmission axis was either parallel (transmits P) or perpendicular (transmits S component)
to the plane of incidence. The LEDs suOpen asset ↗GitHub · plantPhotobiologyLab/openRayTracerpdf-raw-page:6 lines:1-45Code / dataset availability confirmedEurope PMC · checked 7 Sept 2026
Hyperspectral imaging combined with chemometric approaches is proven to be a powerful tool for the quality evaluation and control of fruits. In fruit defect-detection scenarios, developing an unsupervised anomaly detection framework is vital, as defect sample preparation is labor-intensive and time-consuming, especially for exploring potential defects. In this paper, a spectral-spatial, information-based, self-supervised anomaly detection (SSAD) approach is proposed. During training, an auxiliary classifier is proposed to identify the projection axes of principal component (PC) images that were transformed from the hyperspectral data cubes. In test time, the fully connected layer of the learned classifier was used as a 'spectral-spatial' feature extractor, and the feature similarity metric was adopted as the score function for the downstream anomaly evaluation task. The proposed network was evaluated with two fruit data sets: a strawberry data set with bruised, infected, chilling-injured, and contaminated test samples and a blueberry data set with bruised, infected, chilling-injured, and wrinkled samples as anomalies. The results show that the SSAD yielded the best anomaly detection performance (AUC = 0.923 on average) over the baseline methods, and the visualization results further confirmed its advantage in extracting effective 'spectral-spatial' latent representation. Moreover, the robustness of SSAD is verified with the data pollution experiment; it performed significantly better than the baselines when a portion of anomalous samples was involved in the training process.
Why it matches plant phenotyping methods果実の病害・損傷・低温障害などの状態をハイパースペクトル画像から検出する手法を開発・評価しており、植物器官の状態推定が研究の中心です。
abstractIn this paper, a spectral-spatial, information-based, self-supervised anomaly detection (SSAD) approach is proposed.
Reproduction assets foundThe paper's SSAD code implementation and learned models are publicly available on GitHub. The fruit hyperspectral datasets are paper-specific but only available on request from the corresponding author.Code · publicThe code implementation and learned models of SSAD are available at https://github.com/YisenLiu-Intelligent-Sensing/SSAD accessed on 18 May 2022.Open asset ↗YisenLiu-Intelligent-Sensing/SSADlines:57-72Code / dataset availability confirmedEurope PMC · checked 8 Sept 2026
Published1 Feb 2022Proceedings of the National Academy of Sciences of the United States of AmericaCited by 220 · OpenAlex ↗
Although they are staple foods in cuisines globally, many commercial fruit varieties have become progressively less flavorful over time. Due to the cost and difficulty associated with flavor phenotyping, breeding programs have long been challenged in selecting for this complex trait. To address this issue, we leveraged targeted metabolomics of diverse tomato and blueberry accessions and their corresponding consumer panel ratings to create statistical and machine learning models that can predict sensory perceptions of fruit flavor. Using these models, a breeding program can assess flavor ratings for a large number of genotypes, previously limited by the low throughput of consumer sensory panels. The ability to predict consumer ratings of liking, sweet, sour, umami, and flavor intensity was evaluated by a 10-fold cross-validation, and the accuracies of 18 different models were assessed. The prediction accuracies were high for most attributes and ranged from 0.87 for sourness intensity in blueberry using XGBoost to 0.46 for overall liking in tomato using linear regression. Further, the best-performing models were used to infer the flavor compounds (sugars, acids, and volatiles) that contribute most to each flavor attribute. We found that the variance decomposition of overall liking score estimates that 42% and 56% of the variance was explained by volatile organic compounds in tomato and blueberry, respectively. We expect that these models will enable an earlier incorporation of flavor as breeding targets and encourage selection and release of more flavorful fruit varieties.
Why it matches plant phenotyping methods果実の風味という植物器官形質を、メタボロームから予測する統計・機械学習モデルを開発し、交差検証で性能評価している。育種に利用可能な風味表現型推定法が中心であり、単なる代謝測定ではない。
abstractwe leveraged targeted metabolomics of diverse tomato and blueberry accessions and their corresponding consumer panel ratings to create statistical and machine learning models that can predict sensory perceptions of fruit flavor.
Reproduction assets foundThe paper provides public author analysis scripts on GitHub and paper-specific phenotype data (sensory panel ratings, metabolite concentrations, model accuracies) in Datasets S1–S7 within the PNAS supporting information. The caret R package is a generic library and excluded.Code · publicRelevant scripts are provided in the GitHub repository at https://github.com/Resende-Lab/metabolomic_selection_for_enhanced_fruit_flavor .Open asset ↗Resende-Lab/metabolomic_selection_for_enhanced_fruit_flavorlines:126-357Dataset · publicSensory panel ratings and metabolite concentrations are provided in Datasets S1 and S2 . Underlying data for Fig. 3 are provided in Dataset S3 . Model accuracies in Fig. 4 are provided in Datasets S4–S7 .Open asset ↗lines:126-357