Refine
Has Fulltext
- yes (54)
Is part of the Bibliography
- yes (54)
Document Type
- Journal article (51)
- Doctoral Thesis (3)
Keywords
- machine learning (54) (remove)
Institute
- Institut für Geographie und Geologie (10)
- Institut für Informatik (9)
- Center for Computational and Theoretical Biology (5)
- Institut für Klinische Epidemiologie und Biometrie (5)
- Betriebswirtschaftliches Institut (4)
- Medizinische Klinik und Poliklinik II (4)
- Pathologisches Institut (4)
- Theodor-Boveri-Institut für Biowissenschaften (4)
- Institut für diagnostische und interventionelle Radiologie (Institut für Röntgendiagnostik) (3)
- Klinik und Poliklinik für Dermatologie, Venerologie und Allergologie (3)
Sonstige beteiligte Institutionen
An approach to aerodynamically optimizing cycling posture and reducing drag in an Ironman (IM) event was elaborated. Therefore, four commonly used positions in cycling were investigated and simulated for a flow velocity of 10 m/s and yaw angles of 0–20° using OpenFoam-based Nabla Flow CFD simulation software software. A cyclist was scanned using an IPhone 12, and a special-purpose meshing software BLENDER was used. Significant differences were observed by changing and optimizing the cyclist’s posture. Aerodynamic drag coefficient (CdA) varies by more than a factor of 2, ranging from 0.214 to 0.450. Within a position, the CdA tends to increase slightly at yaw angles of 5–10° and decrease at higher yaw angles compared to a straight head wind, except for the time trial (TT) position. The results were applied to the IM Hawaii bike course (180 km), estimating a constant power output of 300 W. Including the wind distributions, two different bike split models for performance prediction were applied. Significant time saving of roughly 1 h was found. Finally, a machine learning approach to deduce 3D triangulation for specific body shapes from 2D pictures was tested.
Neural networks have to capture mathematical relationships in order to learn various tasks. They approximate these relations implicitly and therefore often do not generalize well. The recently proposed Neural Arithmetic Logic Unit (NALU) is a novel neural architecture which is able to explicitly represent the mathematical relationships by the units of the network to learn operations such as summation, subtraction or multiplication. Although NALUs have been shown to perform well on various downstream tasks, an in-depth analysis reveals practical shortcomings by design, such as the inability to multiply or divide negative input values or training stability issues for deeper networks. We address these issues and propose an improved model architecture. We evaluate our model empirically in various settings from learning basic arithmetic operations to more complex functions. Our experiments indicate that our model solves stability issues and outperforms the original NALU model in means of arithmetic precision and convergence.
Objectives
Embedded in the Collaborative Research Center “Fear, Anxiety, Anxiety Disorders” (CRC‐TRR58), this bicentric clinical study aims at identifying biobehavioral markers of treatment (non‐)response by applying machine learning methodology with an external cross‐validation protocol. We hypothesize that a priori prediction of treatment (non‐)response is possible in a second, independent sample based on multimodal markers.
Methods
One‐session virtual reality exposure treatment (VRET) with patients with spider phobia was conducted on two sites. Clinical, neuroimaging, and genetic data were assessed at baseline, post‐treatment and after 6 months. The primary and secondary outcomes defining treatment response are as follows: 30% reduction regarding the individual score in the Spider Phobia Questionnaire and 50% reduction regarding the individual distance in the behavioral avoidance test.
Results
N = 204 patients have been included (n = 100 in Würzburg, n = 104 in Münster). Sample characteristics for both sites are comparable.
Discussion
This study will offer cross‐validated theranostic markers for predicting the individual success of exposure‐based therapy. Findings will support clinical decision‐making on personalized therapy, bridge the gap between basic and clinical research, and bring stratified therapy into reach. The study is registered at ClinicalTrials.gov (ID: NCT03208400).
The identification of biomarker signatures is important for cancer diagnosis and prognosis. However, the detection of clinical reliable signatures is influenced by limited data availability, which may restrict statistical power. Moreover, methods for integration of large sample cohorts and signature identification are limited. We present a step-by-step computational protocol for functional gene expression analysis and the identification of diagnostic and prognostic signatures by combining meta-analysis with machine learning and survival analysis. The novelty of the toolbox lies in its all-in-one functionality, generic design, and modularity. It is exemplified for lung cancer, including a comprehensive evaluation using different validation strategies. However, the protocol is not restricted to specific disease types and can therefore be used by a broad community. The accompanying R package vignette runs in ~1 h and describes the workflow in detail for use by researchers with limited bioinformatics training.
Background
Medical resource management can be improved by assessing the likelihood of prolonged length of stay (LOS) for head and neck cancer surgery patients. The objective of this study was to develop predictive models that could be used to determine whether a patient's LOS after cancer surgery falls within the normal range of the cohort.
Methods
We conducted a retrospective analysis of a dataset consisting of 300 consecutive patients who underwent head and neck cancer surgery between 2017 and 2022 at a single university medical center. Prolonged LOS was defined as LOS exceeding the 75th percentile of the cohort. Feature importance analysis was performed to evaluate the most important predictors for prolonged LOS. We then constructed 7 machine learning and deep learning algorithms for the prediction modeling of prolonged LOS.
Results
The algorithms reached accuracy values of 75.40 (radial basis function neural network) to 97.92 (Random Trees) for the training set and 64.90 (multilayer perceptron neural network) to 84.14 (Random Trees) for the testing set. The leading parameters predicting prolonged LOS were operation time, ischemia time, the graft used, the ASA score, the intensive care stay, and the pathological stages. The results revealed that patients who had a higher number of harvested lymph nodes (LN) had a lower probability of recurrence but also a greater LOS. However, patients with prolonged LOS were also at greater risk of recurrence, particularly when fewer (LN) were extracted. Further, LOS was more strongly correlated with the overall number of extracted lymph nodes than with the number of positive lymph nodes or the ratio of positive to overall extracted lymph nodes, indicating that particularly unnecessary lymph node extraction might be associated with prolonged LOS.
Conclusions
The results emphasize the need for a closer follow-up of patients who experience prolonged LOS. Prospective trials are warranted to validate the present results.
Background: Oro-antral communication (OAC) is a common complication following the extraction of upper molar teeth. The Archer and the Root Sinus (RS) systems can be used to classify impacted teeth in panoramic radiographs. The Archer classes B-D and the Root Sinus classes III, IV have been associated with an increased risk of OAC following tooth extraction in the upper molar region. In our previous study, we found that panoramic radiographs are not reliable for predicting OAC. This study aimed to (1) determine the feasibility of automating the classification (Archer/RS classes) of impacted teeth from panoramic radiographs, (2) determine the distribution of OAC stratified by classification system classes for the purposes of decision tree construction, and (3) determine the feasibility of automating the prediction of OAC utilizing the mentioned classification systems. Methods: We utilized multiple supervised pre-trained machine learning models (VGG16, ResNet50, Inceptionv3, EfficientNet, MobileNetV2), one custom-made convolutional neural network (CNN) model, and a Bag of Visual Words (BoVW) technique to evaluate the performance to predict the clinical classification systems RS and Archer from panoramic radiographs (Aim 1). We then used Chi-square Automatic Interaction Detectors (CHAID) to determine the distribution of OAC stratified by the Archer/RS classes to introduce a decision tree for simple use in clinics (Aim 2). Lastly, we tested the ability of a multilayer perceptron artificial neural network (MLP) and a radial basis function neural network (RBNN) to predict OAC based on the high-risk classes RS III, IV, and Archer B-D (Aim 3). Results: We achieved accuracies of up to 0.771 for EfficientNet and MobileNetV2 when examining the Archer classification. For the AUC, we obtained values of up to 0.902 for our custom-made CNN. In comparison, the detection of the RS classification achieved accuracies of up to 0.792 for the BoVW and an AUC of up to 0.716 for our custom-made CNN. Overall, the Archer classification was detected more reliably than the RS classification when considering all algorithms. CHAID predicted 77.4% correctness for the Archer classification and 81.4% for the RS classification. MLP (AUC: 0.590) and RBNN (AUC: 0.590) for the Archer classification as well as MLP 0.638) and RBNN (0.630) for the RS classification did not show sufficient predictive capability for OAC. Conclusions: The results reveal that impacted teeth can be classified using panoramic radiographs (best AUC: 0.902), and the classification systems can be stratified according to their relationship to OAC (81.4% correct for RS classification). However, the Archer and RS classes did not achieve satisfactory AUCs for predicting OAC (best AUC: 0.638). Additional research is needed to validate the results externally and to develop a reliable risk stratification tool based on the present findings.
Associations between periodontitis and COPD: An artificial intelligence-based analysis of NHANES III
(2022)
A number of cross-sectional epidemiological studies suggest that poor oral health is associated with respiratory diseases. However, the number of cases within the studies was limited, and the studies had different measurement conditions. By analyzing data from the National Health and Nutrition Examination Survey III (NHANES III), this study aimed to investigate possible associations between chronic obstructive pulmonary disease (COPD) and periodontitis in the general population. COPD was diagnosed in cases where FEV (1)/FVC ratio was below 70% (non-COPD versus COPD; binary classification task). We used unsupervised learning utilizing k-means clustering to identify clusters in the data. COPD classes were predicted with logistic regression, a random forest classifier, a stochastic gradient descent (SGD) classifier, k-nearest neighbors, a decision tree classifier, Gaussian naive Bayes (GaussianNB), support vector machines (SVM), a custom-made convolutional neural network (CNN), a multilayer perceptron artificial neural network (MLP), and a radial basis function neural network (RBNN) in Python. We calculated the accuracy of the prediction and the area under the curve (AUC). The most important predictors were determined using feature importance analysis. Results: Overall, 15,868 participants and 19 feature variables were included. Based on k-means clustering, the data were separated into two clusters that identified two risk characteristic groups of patients. The algorithms reached AUCs between 0.608 (DTC) and 0.953% (CNN) for the classification of COPD classes. Feature importance analysis of deep learning algorithms indicated that age and mean attachment loss were the most important features in predicting COPD. Conclusions: Data analysis of a large population showed that machine learning and deep learning algorithms could predict COPD cases based on demographics and oral health feature variables. This study indicates that periodontitis might be an important predictor of COPD. Further prospective studies examining the association between periodontitis and COPD are warranted to validate the present results.
In the past decades, various Earth observation-based time series products have emerged, which have enabled studies and analysis of global change processes. Besides their contribution to understanding past processes, time series datasets hold enormous potential for predictive modeling and thereby meet the demands of decision makers on future scenarios. In order to further exploit these data, a novel pixel-based approach has been introduced, which is the spatio-temporal matrix (STM). The approach integrates the historical characteristics of a specific land cover at a high temporal frequency in order to interpret the spatial and temporal information for the neighborhood of a given target pixel. The provided information can be exploited with common predictive models and algorithms. In this study, this approach was utilized and evaluated for the prediction of future urban/built-settlement growth. Random forest and multi-layer perceptron were employed for the prediction. The tests have been carried out with training strategies based on a one-year and a ten-year time span for the urban agglomerations of Surat (India), Ho-Chi-Minh City (Vietnam), and Abidjan (Ivory Coast). The slope, land use, exclusion, urban, transportation, hillshade (SLEUTH) model was selected as a baseline indicator for the performance evaluation. The statistical results from the receiver operating characteristic curve (ROC) demonstrate a good ability of the STM to facilitate the prediction of future settlement growth and its transferability to different cities, with area under the curve (AUC) values greater than 0.85. Compared with SLEUTH, the STM-based model achieved higher AUC in all of the test cases, while being independent of the additional datasets for the restricted and the preferential development areas.
Highlights
• Brain connectivity states identified by cofluctuation strength.
• CMEP as new method to robustly predict human traits from brain imaging data.
• Network-identifying connectivity ‘events’ are not predictive of cognitive ability.
• Sixteen temporally independent fMRI time frames allow for significant prediction.
• Neuroimaging-based assessment of cognitive ability requires sufficient scan lengths.
Abstract
Human functional brain connectivity can be temporally decomposed into states of high and low cofluctuation, defined as coactivation of brain regions over time. Rare states of particularly high cofluctuation have been shown to reflect fundamentals of intrinsic functional network architecture and to be highly subject-specific. However, it is unclear whether such network-defining states also contribute to individual variations in cognitive abilities – which strongly rely on the interactions among distributed brain regions. By introducing CMEP, a new eigenvector-based prediction framework, we show that as few as 16 temporally separated time frames (< 1.5% of 10 min resting-state fMRI) can significantly predict individual differences in intelligence (N = 263, p < .001). Against previous expectations, individual's network-defining time frames of particularly high cofluctuation do not predict intelligence. Multiple functional brain networks contribute to the prediction, and all results replicate in an independent sample (N = 831). Our results suggest that although fundamentals of person-specific functional connectomes can be derived from few time frames of highest connectivity, temporally distributed information is necessary to extract information about cognitive abilities. This information is not restricted to specific connectivity states, like network-defining high-cofluctuation states, but rather reflected across the entire length of the brain connectivity time series.
Grünflächen stellen einen der wichtigsten Umwelteinflüsse in der Wohnumwelt der Menschen dar. Einerseits wirken sie sich positiv auf die physische und mentale Gesundheit der Menschen aus, andererseits können Grünflächen auch negative Wirkungen anderer Faktoren abmildern, wie beispielsweise die im Laufe des Klimawandels zunehmenden Hitzeereignisse. Dennoch sind Grünflächen nicht für die gesamte Bevölkerung gleichermaßen zugänglich. Bestehende Forschung im Kontext der Umweltgerechtigkeit (UG) konnte bereits aufzeigen, dass unterschiedliche sozio-ökonomische und demographische Gruppen der deutschen Bevölkerung unterschiedlichen Zugriff auf Grünflächen haben. An bestehenden Analysen von Umwelteinflüssen im Kontext der UG wird kritisiert, dass die Auswertung geographischer Daten häufig auf zu stark aggregiertem Level geschieht, wodurch lokal spezifische Expositionen nicht mehr genau abgebildet werden. Dies trifft insbesondere für großflächig angelegte Studien zu. So werden wichtige räumliche Informationen verloren. Doch moderne Erdbeobachtungs- und Geodaten sind so detailliert wie nie und Methoden des maschinellen Lernens ermöglichen die effiziente Verarbeitung zur Ableitung höherwertiger Informationen.
Das übergeordnete Ziel dieser Arbeit besteht darin, am Beispiel von Grünflächen in Deutschland methodische Schritte der systematischen Umwandlung umfassender Geodaten in relevante Geoinformationen für die großflächige und hochaufgelöste Analyse von Umwelteigenschaften aufzuzeigen und durchzuführen. An der Schnittstelle der Disziplinen Fernerkundung, Geoinformatik, Sozialgeographie und Umweltgerechtigkeitsforschung sollen Potenziale moderner Methoden für die Verbesserung der räumlichen und semantischen Auflösung von Geoinformationen erforscht werden. Hierfür werden Methoden des maschinellen Lernens eingesetzt, um Landbedeckung und -nutzung auf nationaler Ebene zu erfassen. Diese Entwicklungen sollen dazu beitragen bestehende Datenlücken zu schließen und Aufschluss über die Verteilungsgerechtigkeit von Grünflächen zu bieten.
Diese Dissertation gliedert sich in drei konzeptionelle Teilschritte. Im ersten Studienteil werden Erdbeobachtungsdaten der Sentinel-2 Satelliten zur deutschlandweiten Klassifikation von Landbedeckungsinformationen verwendet. In Kombination mit punktuellen Referenzdaten der europaweiten Erfassung für Landbedeckungs- und Landnutzungsinformationen des Land Use and Coverage Area Frame Survey (LUCAS) wird ein maschinelles Lernverfahren trainiert. In diesem Kontext werden verschiedene Vorverarbeitungsschritte der LUCAS-Daten und deren Einfluss auf die Klassifikationsgenauigkeit beleuchtet. Das Klassifikationsverfahren ist in der Lage Landbedeckungsinformationen auch in komplexen urbanen Gebieten mit hoher Genauigkeit abzuleiten. Ein Ergebnis des Studienteils ist eine deutschlandweite Landbedeckungsklassifikation mit einer Gesamtgenauigkeit von 93,07 %, welche im weiteren Verlauf der Arbeit genutzt wird, um grüne Landbedeckung (GLC) räumlich zu quantifizieren.
Im zweiten konzeptionellen Teil der Arbeit steht die differenzierte Betrachtung von Grünflächen anhand des Beispiels öffentlicher Grünflächen (PGS), die häufig Gegenstand der UG-Forschung ist, im Vordergrund. Doch eine häufig verwendete Quelle für räumliche Daten zu öffentlichen Grünflächen, der European Urban Atlas (EUA), wird bisher nicht flächendeckend für Deutschland erhoben. Dieser Studienteil verfolgt einen datengetriebenen Ansatz, die Verfügbarkeit von öffentlichem Grün auf der räumlichen Ebene von Nachbarschaften für ganz Deutschland zu ermitteln. Hierfür dienen bereits vom EUA erfasste Gebiete als Referenz. Mithilfe einer Kombination von Erdbeobachtungsdaten und Informationen aus dem OpenStreetMap-Projekt wird ein Deep Learning -basiertes Fusionsnetzwerk erstellt, welche die verfügbare Fläche von öffentlichem Grün quantifiziert. Das Ergebnis dieses Schrittes ist ein Modell, welches genutzt wird, um die Menge öffentlicher Grünflächen in der Nachbarschaft zu schätzen (𝑅 2 = 0.952).
Der dritte Studienteil greift die Ergebnisse der ersten beiden Studienteile auf und betrachtet die Verteilung von Grünflächen in Deutschland unter Hinzunahme von georeferenzierten Bevölkerungsdaten. Diese exemplarische Analyse unterscheidet dabei Grünflächen nach zwei Typen: GLC und PGS. Zunächst wird mithilfe deskriptiver Statistiken die generelle Grünflächenverteilung in der Bevölkerung Deutschlands beleuchtet. Daraufhin wird die Verteilungsgerechtigkeit anhand gängiger Gerechtigkeitsmetriken bestimmt. Abschließend werden die Zusammenhänge zwischen der demographischen Komposition der Nachbarschaft und der verfügbaren Menge von Grünflächen anhand dreier exemplarischer soziodemographischer Gesellschaftsgruppen untersucht. Die Analyse zeigt starke Unterschiede der Verfügbarkeit von PGS zwischen städtischen und ländlichen Gebieten. Ein höherer Prozentsatz der Stadtbevölkerung hat Zugriff das Mindestmaß von PGS gemessen an der Vorgabe der Weltgesundheitsorganisation. Die Ergebnisse zeigen auch einen deutlichen Unterschied bezüglich der Verteilungsgerechtigkeit zwischen GLC und PGS und verdeutlichen die Relevanz der Unterscheidung von Grünflächentypen für derartige
Untersuchungen. Die abschließende Betrachtung verschiedener Bevölkerungsgruppen arbeitet Unterschiede auf soziodemographischer Ebene auf.
In der Zusammenschau demonstriert diese Arbeit wie moderne Geodaten und Methoden des maschinellen Lernens genutzt werden können bisherige Limitierungen räumlicher Datensätze zu überwinden. Am Beispiel von Grünflächen in der Wohnumgebung der Bevölkerung Deutschlands wird gezeigt, dass landesweite Analysen zur Umweltgerechtigkeit durch hochaufgelöste und lokal feingliedrige geographische Informationen bereichert werden können. Diese Arbeit verdeutlicht, wie die Methoden der Erdbeobachtung und Geoinformatik einen wichtigen Beitrag leisten können, die Ungleichheit der Wohnumwelt der Menschen zu identifizieren und schlussendlich den nachhaltigen Siedlungsbau in Form von objektiven Informationen zu unterstützen und überwachen.
Deep convolutional generative adversarial networks (GAN) allow for creating images from existing databases. We applied a modified light-weight GAN (FastGAN) algorithm to cerebral blood flow SPECTs and aimed to evaluate whether this technology can generate created images close to real patients. Investigating three anatomical levels (cerebellum, CER; basal ganglia, BG; cortex, COR), 551 normal (248 CER, 174 BG, 129 COR) and 387 pathological brain SPECTs using N-isopropyl p-I-123-iodoamphetamine (123I-IMP) were included. For the latter scans, cerebral ischemic disease comprised 291 uni- (66 CER, 116 BG, 109 COR) and 96 bilateral defect patterns (44 BG, 52 COR). Our model was trained using a three-compartment anatomical input (dataset ‘A’; including CER, BG, and COR), while for dataset ‘B’, only one anatomical region (COR) was included. Quantitative analyses provided mean counts (MC) and left/right (LR) hemisphere ratios, which were then compared to quantification from real images. For MC, ‘B’ was significantly different for normal and bilateral defect patterns (P < 0.0001, respectively), but not for unilateral ischemia (P = 0.77). Comparable results were recorded for LR, as normal and ischemia scans were significantly different relative to images acquired from real patients (P ≤ 0.01, respectively). Images provided by ‘A’, however, revealed comparable quantitative results when compared to real images, including normal (P = 0.8) and pathological scans (unilateral, P = 0.99; bilateral, P = 0.68) for MC. For LR, only uni- (P = 0.03), but not normal or bilateral defect scans (P ≥ 0.08) reached significance relative to images of real patients. With a minimum of only three anatomical compartments serving as stimuli, created cerebral SPECTs are indistinguishable to images from real patients. The applied FastGAN algorithm may allow to provide sufficient scan numbers in various clinical scenarios, e.g., for “data-hungry” deep learning technologies or in the context of orphan diseases.
Purpose
Machine learning based on radiomics features has seen huge success in a variety of clinical applications. However, the need for standardization and reproducibility has been increasingly recognized as a necessary step for future clinical translation. We developed a novel, intuitive open-source framework to facilitate all data analysis steps of a radiomics workflow in an easy and reproducible manner and evaluated it by reproducing classification results in eight available open-source datasets from different clinical entities.
Methods
The framework performs image preprocessing, feature extraction, feature selection, modeling, and model evaluation, and can automatically choose the optimal parameters for a given task. All analysis steps can be reproduced with a web application, which offers an interactive user interface and does not require programming skills. We evaluated our method in seven different clinical applications using eight public datasets: six datasets from the recently published WORC database, and two prostate MRI datasets—Prostate MRI and Ultrasound With Pathology and Coordinates of Tracked Biopsy (Prostate-UCLA) and PROSTATEx.
Results
In the analyzed datasets, AutoRadiomics successfully created and optimized models using radiomics features. For WORC datasets, we achieved AUCs ranging from 0.56 for lung melanoma metastases detection to 0.93 for liposarcoma detection and thereby managed to replicate the previously reported results. No significant overfitting between training and test sets was observed. For the prostate cancer detection task, results were better in the PROSTATEx dataset (AUC = 0.73 for prostate and 0.72 for lesion mask) than in the Prostate-UCLA dataset (AUC 0.61 for prostate and 0.65 for lesion mask), with external validation results varying from AUC = 0.51 to AUC = 0.77.
Conclusion
AutoRadiomics is a robust tool for radiomic studies, which can be used as a comprehensive solution, one of the analysis steps, or an exploratory tool. Its wide applicability was confirmed by the results obtained in the diverse analyzed datasets. The framework, as well as code for this analysis, are publicly available under https://github.com/pwoznicki/AutoRadiomics.
Objectives
Open-access cancer imaging datasets have become integral for evaluating novel AI approaches in radiology. However, their use in quantitative analysis with radiomics features presents unique challenges, such as incomplete documentation, low visibility, non-uniform data formats, data inhomogeneity, and complex preprocessing. These issues may cause problems with reproducibility and standardization in radiomics studies.
Methods
We systematically reviewed imaging datasets with public copyright licenses, published up to March 2023 across four large online cancer imaging archives. We included only datasets with tomographic images (CT, MRI, or PET), segmentations, and clinical annotations, specifically identifying those suitable for radiomics research. Reproducible preprocessing and feature extraction were performed for each dataset to enable their easy reuse.
Results
We discovered 29 datasets with corresponding segmentations and labels in the form of health outcomes, tumor pathology, staging, imaging-based scores, genetic markers, or repeated imaging. We compiled a repository encompassing 10,354 patients and 49,515 scans. Of the 29 datasets, 15 were licensed under Creative Commons licenses, allowing both non-commercial and commercial usage and redistribution, while others featured custom or restricted licenses. Studies spanned from the early 1990s to 2021, with the majority concluding after 2013. Seven different formats were used for the imaging data. Preprocessing and feature extraction were successfully performed for each dataset.
Conclusion
RadiomicsHub is a comprehensive public repository with radiomics features derived from a systematic review of public cancer imaging datasets. By converting all datasets to a standardized format and ensuring reproducible and traceable processing, RadiomicsHub addresses key reproducibility and standardization challenges in radiomics.
Critical relevance statement
This study critically addresses the challenges associated with locating, preprocessing, and extracting quantitative features from open-access datasets, to facilitate more robust and reliable evaluations of radiomics models.
Key points
- Through a systematic review, we identified 29 cancer imaging datasets suitable for radiomics research.
- A public repository with collection overview and radiomics features, encompassing 10,354 patients and 49,515 scans, was compiled.
- Most datasets can be shared, used, and built upon freely under a Creative Commons license.
- All 29 identified datasets have been converted into a common format to enable reproducible radiomics feature extraction.
Among the defense strategies developed in microbes over millions of years, the innate adaptive CRISPR-Cas immune systems have spread across most of bacteria and archaea. The flexibility, simplicity, and specificity of CRISPR-Cas systems have laid the foundation for CRISPR-based genetic tools. Yet, the efficient administration of CRISPR-based tools demands rational designs to maximize the on-target efficiency and off-target specificity. Specifically, the selection of guide RNAs (gRNAs), which play a crucial role in the target recognition of CRISPR-Cas systems, is non-trivial. Despite the fact that the emerging machine learning techniques provide a solution to aid in gRNA design with prediction algorithms, design rules for many CRISPR-Cas systems are ill-defined, hindering their broader applications.
CRISPR interference (CRISPRi), an alternative gene silencing technique using a catalytically dead Cas protein to interfere with transcription, is a leading technique in bacteria for functional interrogation, pathway manipulation, and genome-wide screens. Although the application is promising, it also is hindered by under-investigated design rules. Therefore, in this work, I develop a state-of-art predictive machine learning model for guide silencing efficiency in bacteria leveraging the advantages of feature engineering, data integration, interpretable AI, and automated machine learning. I first systematically investigate the influential factors that attribute to the extent of depletion in multiple CRISPRi genome-wide essentiality screens in Escherichia coli and demonstrate the surprising dominant contribution of gene-specific effects, such as gene expression level. These observations allowed me to segregate the confounding gene-specific effects using a mixed-effect random forest (MERF) model to provide a better estimate of guide efficiency, together with the improvement led by integrating multiple screens. The MERF model outperformed existing tools in an independent high-throughput saturating screen. I next interpret the predictive model to extract the design rules for robust gene silencing, such as the preference for cytosine and disfavoring for guanine and thymine within and around the protospacer adjacent motif (PAM) sequence. I further incorporated the MERF model in a web-based tool that is freely accessible at www.ciao.helmholtz-hiri.de.
When comparing the MERF model with existing tools, the performance of the alternative gRNA design tool optimized for CRISPRi in eukaryotes when applied to bacteria was far from satisfying, questioning the robustness of prediction algorithms across organisms. In addition, the CRISPR-Cas systems exhibit diverse mechanisms albeit with some similarities. The captured predictive patterns from one dataset thereby are at risk of poor generalization when applied across organisms and CRISPR-Cas techniques. To fill the gap, the machine learning approach I present here for CRISPRi could serve as a blueprint for the effective development of prediction algorithms for specific organisms or CRISPR-Cas systems of interest. The explicit workflow includes three principle steps: 1) accommodating the feature set for the CRISPR-Cas system or technique; 2) optimizing a machine learning model using automated machine learning; 3) explaining the model using interpretable AI. To illustrate the applicability of the workflow and diversity of results when applied across different bacteria and CRISPR-Cas systems, I have applied this workflow to analyze three distinct CRISPR-Cas genome-wide screens. From the CRISPR base editor essentiality screen in E. coli, I have determined the PAM preference and sequence context in the editing window for efficient editing, such as A at the 2nd position of PAM, A/TT/TG downstream of PAM, and TC at the 4th to 5th position of gRNAs. From the CRISPR-Cas13a screen in E. coli, in addition to the strong correlation with the guide depletion, the target expression level is the strongest predictor in the model, supporting it as a main determinant of the activation of Cas13-induced immunity and better characterizing the CRISPR-Cas13 system. From the CRISPR-Cas12a screen in Klebsiella pneumoniae, I have extracted the design rules for robust antimicrobial activity across K. pneumoniae strains and provided a predictive algorithm for gRNA design, facilitating CRISPR-Cas12a as an alternative technique to tackle antibiotic resistance.
Overall, this thesis presents an accurate prediction algorithm for CRISPRi guide efficiency in bacteria, providing insights into the determinants of efficient silencing and guide designs. The systematic exploration has led to a robust machine learning approach for effective model development in other bacteria and CRISPR-Cas systems. Applying the approach in the analysis of independent CRISPR-Cas screens not only sheds light on the design rules but also the mechanisms of the CRISPR-Cas systems. Together, I demonstrate that applied machine learning paves the way to a deeper understanding and a broader application of CRISPR-Cas systems.