Machine learning methods for extracting patterns from high-dimensional data are very important in the biological sciences. However, in certain cases, real-world applications cannot confirm the reported prediction performance. One of the main reasons for this is data leakage, which can be seen as the illicit sharing of information between the training data and the test data, resulting in performance estimates that are far better than the performance observed in the intended application scenario. Data leakage can be difficult to detect in biological datasets due to their complex dependencies. With this in mind, we present seven questions that should be asked to prevent data leakage when constructing machine learning models in biological domains. We illustrate the usefulness of our questions by applying them to nontrivial examples. Our goal is to raise awareness of potential data leakage problems and to promote robust and reproducible machine learning-based research in biology.
Mehr
Florian Haselbeck,
Viola Stiele,
Claudia Guadarrama Serrano,
Kelly Lim-Trinh,
Klaus Menrad,
Thomas Hannus,
Dominik Grimm
Accurately forecasting demand is a potential competitive advantage, especially when dealing with perishable products. The multi-billion dollar horticultural industry is highly affected by perishability, but has received limited attention in forecasting research. In this paper, we analyze the applicability of general compared to dataset-specific predictors, as well as the influence of external information and online model update schemes. We employ a heterogeneous set of horticultural data, three classical, and twelve machine learning-based forecasting approaches. Our results show a superiority of multivariate machine learning methods, in particular the ensemble learner XGBoost. These advantages highlight the importance of external factors, with the feature set containing statistical, calendrical, and weather-related features leading to the most robust performance. We further observe that a general model is unable to capture the heterogeneity of the data and is outperformed by dataset-specific predictors. Moreover, frequent model updates have a negligible impact on forecasting quality, allowing long-term forecasting without significant performance degradation.
Mehr
Florian Haselbeck,
Maura John,
Yuqi Zhang,
Juan Pablo Fuenzalida-Werner,
Ruben Costa,
Dominik Grimm
Protein thermostability is important in many areas of biotechnology, including enzyme engineering and protein-hybrid optoelectronics. Ever-growing protein databases and information on stability at different temperatures allow the training of machine learning models to predict whether proteins are thermophilic. In silico predictions could reduce costs and accelerate the development process by guiding researchers to more promising candidates. Existing models for predicting protein thermophilicity rely mainly on features derived from physicochemical properties. Recently, modern protein language models that directly use sequence information have demonstrated superior performance in several tasks. In this study, we evaluate the usefulness of protein language model embeddings for thermophilicity prediction with ProLaTherm, a Protein Language model-based Thermophilicity predictor. ProLaTherm significantly outperforms all feature-, sequence- and literature-based comparison partners on multiple evaluation metrics. In terms of the Matthew’s correlation coefficient, ProLaTherm outperforms the second-best competitor by 18.1% in a nested cross-validation setup. Using proteins from species not overlapping with species from the training data, ProLaTherm outperforms all competitors by at least 9.7%. On these data, it misclassified only one nonthermophilic protein as thermophilic. Furthermore, it correctly identified 97.4% of all thermophilic proteins in our test set with an optimal growth temperature above 70°C.
Summary: Time series forecasting is a research area with applications in various domains, nevertheless without yielding a predominant method so far. We present ForeTiS, a comprehensive and open source Python framework that allows rigorous training, comparison, and analysis of state-of-the-art time series forecasting approaches. Our framework includes fully automated yet configurable data preprocessing and feature engineering. In addition, we use advanced Bayesian optimization for automatic hyperparameter search. ForeTiS is easy to use, even for non-programmers, requiring only a single line of code to apply state-of-the-art time series forecasting. Various prediction models, ranging from classical forecasting approaches to machine learning techniques and deep learning architectures, are already integrated. More importantly, as a key benefit for researchers aiming to develop new forecasting models, ForeTiS is designed to allow for rapid integration and fair benchmarking in a reliable framework. Thus, we provide a powerful framework for both end users and forecasting experts.Availability: ForeTiS is available at https://github.com/grimmlab/ForeTiS. We provide a setup using Docker, as well as a Python package at https://pypi.org/project/ForeTiS/. Extensive online documentation with hands-on tutorials and videos can be found at https://foretis.readthedocs.io.
SummaryPredicting complex traits from genotypic information is a major challenge in various biological domains. With easyPheno, we present a comprehensive Python framework enabling the rigorous training, comparison, and analysis of phenotype predictions for a variety of different models, ranging from common genomic selection approaches over classical machine learning and modern deep learning based techniques. Our framework is easy-to-use, also for non-programming-experts, and includes an automatic hyperparameter search using state-of-the-art Bayesian optimization. Moreover, easyPheno provides various benefits for bioinformaticians developing new prediction models. easyPheno enables to quickly integrate novel models and functionalities in a reliable framework and to benchmark against various integrated prediction models in a comparable setup. In addition, the framework allows the assessment of newly developed prediction models under pre-defined settings using simulated data. We provide a detailed documentation with various hands-on tutorials and videos explaining the usage of easyPheno to novice users.Availability and ImplementationeasyPheno is publicly available at https://github.com/grimmlab/easyPheno and can be easily installed as Python package via https://pypi.org/project/easypheno/ or using Docker.Supplementary informationA comprehensive documentation including various tutorials complemented with videos can be found at https://easypheno.readthedocs.io/. In addition, we provide examples of how to use easyPheno with real and simulated data in the Supplementary.
Mehr
Nikita Genze,
Raymond Ajekwe,
Zeynep Güreli,
Florian Haselbeck,
Michael Grieb,
Dominik Grimm
Weeds are undesired plants in agricultural fields that affect crop yield and quality by competing for nutrients, water, sunlight and space. For centuries, farmers have used several strategies and resources to remove weeds. The use of herbicide is still the most common control strategy. To reduce the amount of herbicide and impact caused by uniform spraying, site-specific weed management (SSWM) through variable rate herbicide application and mechanical weed control have long been recommended. To implement such precise strategies, accurate detection and classification of weeds in crop fields is a crucial first step. Due to the phenotypic similarity between some weeds and crops as well as changing weather conditions, it is challenging to design an automated system for general weed detection. For efficiency, unmanned aerial vehicles (UAV) are commonly used for image capturing. However, high wind pressure and different drone settings have a severe effect on the capturing quality, what potentially results in degraded images, e.g., due to motion blur. In this paper, we investigate the generalization capabilities of Deep Learning methods for early weed detection in sorghum fields under such challenging capturing conditions. For this purpose, we developed weed segmentation models using three different state-of-the-art Deep Learning architectures in combination with residual neural networks as feature extractors.We further publish a manually annotated and expert-curated UAV imagery dataset for weed detection in sorghum fields under challenging conditions. Our results show that our trained models generalize well regarding the detection of weeds, even for degraded captures due to motion blur. An UNet-like architecture with a ResNet-34 feature extractor achieved an F1-score of over 89 % on a hold-out test-set. Further analysis indicate that the trained model performed well in predicting the general plant shape, while most misclassifications appeared at borders of the plants. Beyond that, our approach can detect intra-row weeds without additional information as well as partly occluded plants in contrast to existing research.All data, including the newly generated and annotated UAV imagery dataset, and code is publicly available on GitHub: https://github.com/grimmlab/UAVWeedSegmentation and Mendeley Data: https://doi.org/10.17632/4hh45vkp38.3
Mehr
Maura John,
Florian Haselbeck,
Rupashree Dass,
Christoph Malisi,
Christian Dreischer,
Sebastian J Schultheiss,
Dominik Grimm
Genomic selection is an integral tool for breeders to accurately select plants directly from genotype data leading to faster and more resource-efficient breeding programs. Several prediction methods have been established in the last few years. These range from classical linear mixed models to complex non-linear machine learning approaches, such as Support Vector Regression, and modern deep learning-based architectures. Many of these methods have been extensively evaluated on different crop species with varying outcomes. In this work, our aim is to systematically compare twelve different phenotype prediction models, including basic genomic selection methods to more advanced deep learning-based techniques. More importantly, we assess the performance of these models on simulated phenotype data as well as on real-world data from Arabidopsis thaliana and two breeding datasets from soy and corn. The synthetic phenotypic data allows us to analyze all prediction models and especially the selected markers under controlled and predefined settings. We show that Bayes B and linear regression models with sparsity constraints perform best under different simulation settings with respect to explained variance. Further, we can confirm results from other studies that there is no superiority of more complex neural network-based architectures for phenotype prediction compared to well established methods. However, on real-world data, for which several prediction models yield comparable results with slight advantages for Elastic Net, this picture is less clear, suggesting that there is a lot of room for future research.
Mehr
Florian Haselbeck,
Jennifer Killinger,
Klaus Menrad,
Thomas Hannus,
Dominik Grimm
Forecasting future demand is of high importance for many companies as it affects operational decisions. This is especially relevant for products with a short shelf life due to the potential disposal of unsold items. Horticultural products are highly influenced by this, however with limited attention in forecasting research so far. Beyond that, many forecasting competitions show a competitive performance of classical forecasting methods. For the first time, we empirically compared the performance of nine state-of-the-art machine learning and three classical forecasting algorithms for horticultural sales predictions. We show that machine learning methods were superior in all our experiments, with the gradient boosted ensemble learner XGBoost being the top performer in 14 out of 15 comparisons. This advantage over classical forecasting approaches increased for datasets with multiple seasons. Further, we show that including additional external factors, such as weather and holiday information, as well as meta-features led to a boost in predictive performance. In addition, we investigated whether the algorithms can capture the sudden increase in demand of horticultural products during the SARS-CoV-2 pandemic in 2020. For this special case, XGBoost was also superior. All code and data is publicly available on GitHub: https://github.com/grimmlab/HorticulturalSalesPredictions.
Mehr
Konferenzbeiträge
Georg Back,
Dirk Walther,
Florian Haselbeck
Berechtigungen: Peer Reviewed
Sequence Design for RNA Inverse Folding via Discrete Diffusion Models (2025) German Conference on Bioinformatics 2025 .
Towards Data Efficient, AI-Driven Weed Detection: Leveraging Foundation Models for Diverse Agricultural Conditions (2025) HEF AgriScience Symposium 2025 .
Enhancing Weed Detection with Fine-Tuned Foundation Models for Robust and Generalizable Precision Agriculture (2025) HEFagrar PhD Symposium 2025 .
Georg Back,
Florian Haselbeck
Predictions of DNA mechanical properties at a genomic scale reveal potentially new functional roles of DNA flexibility (2025) Weihenstephan Bioinformatics Symposium 2025 .
Enhancing Weed Detection with Fine-Tuned Foundation Models for Robust and Generalizable Precision Agriculture (2025) Weihenstephan Bioinformatics Symposium 2025 .
Ashima Khanna,
Florian Haselbeck,
Dominik Grimm
Predicting Protein Thermostability through Deep Learning Leveraging 3D Structural Information (2024) Biological Materials Science - A workshop on biogenic, bioinspired, biomimetic and biohybrid materials for innovative optical, photonics and optoelectronics applications 2024 .
In protein engineering, improving thermostability is essential for many industrial and pharmaceutical applications. However, the experimental process of identifying stabilizing mutations is time-consuming due to the enormous search space. With the increasing availability of protein structural and thermostability data, computational approaches using deep learning to identify thermostable candidates are gaining popularity. In this work, we present and benchmark a novel graph neural network, ProtGCN, that incorporates geometric and energetic details of proteins to predict changes in Gibbs free energy (ΔG), a key indicator of thermostability, upon single point mutations. Unlike conventional methods that rely on sequence or structural features, our model uses protein graphs with rich node features, carefully preprocessed from a comprehensive dataset of approximately 4149 mutated sequences across 117 protein families. In addition, ProtGCN is enhanced by incorporating embeddings from the Evolutionary Scale Modeling (ESM) protein language model into the protein graphs. This integration allows ProtGCN (ESM) to outperform comparison models, achieving competitive performance with XGBoost and a protein language model-based multi-layer perceptron on all evaluation metrics, and outperforming all models on further analyses. A strength of ProtGCN (ESM) is its ability to correctly identify and predict stabilizing and destabilizing mutations with extreme effects, which are typically underrepresented in thermostability datasets. These results suggest a promising direction for future computational protein engineering research.
Florian Haselbeck,
Maura John,
Yuqi Zhang,
Juan Pablo Fuenzalida-Werner,
Ruben Costa,
Dominik Grimm
Superior Protein Thermophilicity Prediction With Protein Language Model Embeddings (2024) Biological Materials Science - A workshop on biogenic, bioinspired, biomimetic and biohybrid materials for innovative optical, photonics and optoelectronics applications 2024 .
Protein thermostability is an essential property for many biotechnological fields, such as enzyme engineering and protein-hybrid optoelectronics. In this context, machine learning-based in silico predictions have the potential to reduce costs and development time by identifying the most promising candidates for subsequent experiments. The development of such prediction models is enabled by ever-growing protein databases and information on protein stability at different temperatures. In this study, we leverage protein language model embeddings for thermophilicity prediction with ProLaTherm, a Protein Language model-based Thermophilicity predictor. We assess ProLaTherm against several feature-, sequence-, and literature-based comparison partners on a new benchmark dataset derived from a significant update of published data. ProLaTherm outperforms all comparison partners both in a nested cross-validation setup and on protein sequences from species not seen during training with respect to multiple evaluation metrics. In terms of Matthew's correlation coefficient, ProLaTherm surpasses the second-best competitor by 18.1% in the nested cross-validation setup. Using proteins from species that do not overlap with species from the training data, ProLaTherm outperforms all competitors by at least 9.7%. On this data, it misclassified only one non-thermophilic protein as thermophilic. Furthermore, it correctly identified 97.4% of all thermophilic proteins in our test set with an optimal growth temperature above 70°C.
Jan D Hüwel,
Florian Haselbeck,
Dominik Grimm,
Christian Beecks
One of the major challenges in time series analysis are changing data distributions, especially when processing data streams. To ensure an up-to-date model delivering useful predictions at all times, model reconfigurations are required to adapt to such evolving streams. For Gaussian processes, this might require the adaptation of the internal kernel expression. In this paper, we present dynamically self-adjusting Gaussian processes by introducing Event Triggered Kernel Adjustments in Gaussian process modelling (ETKA), a novel data stream modelling algorithm that can handle evolving and changing data distributions. To this end, we enhance the recently introduced Adjusting Kernel Search with a novel online change point detection method. Our experiments on simulated data with varying change point patterns suggest a broad applicability of ETKA. On real-world data, ETKA outperforms comparison partners that differ regarding the model adjustment and its refitting trigger in nine respective ten out of 14 cases. These results confirm ETKA's ability to enable a more accurate and, in some settings, also more efficient data stream processing via Gaussian processes.Code availability: https://github.com/JanHuewel/ETKA
Timeseriesforecastingisagrowingdomainwithdiverseapplications. However, changes of the system behavior over time due to internal or external influences are challenging. Therefore, predictions of a previously learned forecast- ing model might not be useful anymore. In this paper, we present EVent-triggered Augmented Refitting of Gaussian Process Regression for Seasonal Data (EVARS- GPR), a novel online algorithm that is able to handle sudden shifts in the target variable scale of seasonal data. For this purpose, EVARS-GPR combines online change point detection with a refitting of the prediction model using data aug- mentation for samples prior to a change point. Our experiments on simulated data show that EVARS-GPR is applicable for a wide range of output scale changes. EVARS-GPR has on average a 20.8% lower RMSE on different real-world datasets compared to methods with a similar computational resource consumption. Fur- thermore, we show that our algorithm leads to a six-fold reduction of the averaged runtime in relation to all comparison partners with a periodical refitting strategy. In summary, we present a computationally efficient online forecasting algorithm for seasonal time series with changes of the target variable scale and demonstrate its functionality on simulated as well as real-world data. All code is publicly available on GitHub: https://github.com/grimmlab/evars-gpr.
Time series forecasting is a growing domain with diverse applications. However, changes of the system behavior over time due to internal or external influences are challenging. Therefore, predictions of a previously learned fore-casting model might not be useful anymore. In this paper, we present EVent-triggered Augmented Refitting of Gaussian Process Regression for Seasonal Data (EVARS-GPR), a novel online algorithm that is able to handle sudden shifts in the target variable scale of seasonal data. For this purpose, EVARS-GPR com-bines online change point detection with a refitting of the prediction model using data augmentation for samples prior to a change point. Our experiments on sim-ulated data show that EVARS-GPR is applicable for a wide range of output scale changes. EVARS-GPR has on average a 20.8 % lower RMSE on different real-world datasets compared to methods with a similar computational resource con-sumption. Furthermore, we show that our algorithm leads to a six-fold reduction of the averaged runtime in relation to all comparison partners with a periodical refitting strategy. In summary, we present a computationally efficient online fore-casting algorithm for seasonal time series with changes of the target variable scale and demonstrate its functionality on simulated as well as real-world data. All code is publicly available on GitHub: this https URL
Mehr
Vorträge
Florian Haselbeck
Deep-Learning-gestützte Unkrauterkennung: Aktueller Stand und zukünftige Herausforderungen (2025) HEF AgriScience Symposium 2025 .
Florian Haselbeck
Deep learning-based weed detection: state of the art and future challenges (2025) AgScience-on-Tap 2025 .
One of the major challenges in time series analysis are changing data distributions, especially when processing data streams. To ensure an up-to-date model delivering useful predictions at all times, model reconfigurations are required to adapt to such evolving streams. For Gaussian processes, this might require the adaptation of the internal kernel expression. In this paper, we present dynamically self-adjusting Gaussian processes by introducing Event Triggered Kernel Adjustments in Gaussian process modelling (ETKA), a novel data stream modelling algorithm that can handle evolving and changing data distributions. To this end, we enhance the recently introduced Adjusting Kernel Search with a novel online change point detection method. Our experiments on simulated data with varying change point patterns suggest a broad applicability of ETKA. On real-world data, ETKA outperforms comparison partners that differ regarding the model adjustment and its refitting trigger in nine respective ten out of 14 cases. These results confirm ETKA's ability to enable a more accurate and, in some settings, also more efficient data stream processing via Gaussian processes.Code availability: https://github.com/JanHuewel/ETKA
Mehr
Florian Haselbeck,
Dominik Grimm
EVARS-GPR: EVent-triggered Augmented Refitting of Gaussian Process Regression for Seasonal Data (2021) 44th German Conference on Artificial Intelligence (Virtual Conference) .
DOI: 10.1007/978-3-030-87626-5_11
Time series forecasting is a growing domain with diverse applications. However, changes of the system behavior over time due to internal or external influences are challenging. Therefore, predictions of a previously learned forecasting model might not be useful anymore. In this paper, we present EVent-triggered Augmented Refitting of Gaussian Process Regression for Seasonal Data (EVARS-GPR), a novel online algorithm that is able to handle sudden shifts in the target variable scale of seasonal data. For this purpose, EVARS-GPR combines online change point detection with a refitting of the prediction model using data augmentation for samples prior to a change point. Our experiments on simulated data show that EVARS-GPR is applicable for a wide range of output scale changes. EVARS-GPR has on average a 20.8% lower RMSE on different real-world datasets compared to methods with a similar computational resource consumption. Furthermore, we show that our algorithm leads to a six-fold reduction of the averaged runtime in relation to all comparison partners with a periodical refitting strategy. In summary, we present a computationally efficient online forecasting algorithm for seasonal time series with changes of the target variable scale and demonstrate its functionality on simulated as well as real-world data. All code is publicly available on GitHub: https://github.com/grimmlab/evars-gpr.
Klaus Menrad,
Florian Haselbeck,
Thomas Decker,
Dominik Grimm,
Thomas Hannus,
Kai Sparke,
M. Lehberger,
M. Drechsler,
Andreas Holzapfel,
S. Schröder,
Gerald Neu,
F. Bertlich
Das Projekt adressiert Herausforderungen intakter Waldökosysteme und entwickelt ein wissenschaftlich fundiertes und praxistaugliches Konzept zur klimaresilienten Wiederbewaldung.
…
Wir verwenden Cookies. Einige sind notwendig für die Funktion der Webseite, andere helfen uns, die Webseite zu verbessern. Um unseren eigenen Ansprüchen beim Datenschutz gerecht zu werden, erfassen wir lediglich anonymisierte Nutzerdaten mit „Matomo“. Um unser Internetangebot für Sie ansprechender zu gestalten, binden wir außerdem externe Inhalte unserer Social-Media-Kanäle ein.
