Neighborhood Size Research Articles

Genetic diversity and structure of a recent fish invasion: Tench (Tinca tinca) in eastern North America.

Introduced and geographically expanding populations experience similar eco-evolutionary challenges, including founder events, genetic bottlenecks, and novel environments. Theory predicts that reduced genetic diversity resulting from such phenomena limits the success of introduced populations. Using 1900 SNPs obtained from restriction-site-associated DNA sequencing, we evaluated hypotheses related to the invasion history and connectivity of an invasive population of Tench (Tinca tinca), a Eurasian freshwater fish that has been expanding geographically in eastern North America for three decades. Consistent with the reported history of a single introduction event, our findings suggest that multiple introductions from distinct genetic sources are unlikely as Tench had a small effective population size (~114 [95% CI=106-123] individuals), no strong population subdivision across time and space, and evidence of a recent genetic bottleneck. The large genetic neighbourhood size (220 km) and weak within-population genetic substructure suggested high connectivity across the invaded range, despite the relatively large area occupied. There was some evidence for a small decay in genetic diversity as the species expanded northward, but not southward, into new habitats. As eradicating the species within a ~112 km radius would be necessary to prevent recolonization, eradicating Tench is likely not feasible at watershed-and possibly local-scales. Management should instead focus on reducing abundance in priority conservation areas to mitigate adverse impacts. Our study indicates that introduced populations can thrive and exhibit relatively high levels of genetic diversity despite severe bottlenecks (<1.5% of the ancestral effective population size) and suggests that landscape heterogeneity and population demographics can generate variability in spatial patterns of genetic diversity within a single range expansion.

Social Determinants of Health and Symptom Burden During Cancer Treatment.

Cancer survivors (defined as individuals from diagnosis to the end of life) in treatment experience multiple physical and psychological symptoms (e.g., fatigue, pain, depression, anxiety, disturbed sleep) that influence their well-being and treatment outcomes. Underrepresented cancer survivors may disproportionately experience greater symptom burden (number of symptoms, symptom severity, depression, anxiety). The aim of this study was to examine the relationships of social determinants of health, including age, ethnicity, education, income and whether income meets the survivor's needs, neighborhood (rural vs. urban), access to healthcare (e.g., insurance), and social isolation, with symptom burden in cancer survivors. This secondary analysis included baseline data from 400 cancer survivors of solid tumor cancers undergoing chemotherapy or targeted therapy who participated in a larger randomized trial of symptom management interventions. Symptom burden was measured by the Center for Epidemiological Studies-Depression scale for depression and Patient-Reported Outcomes Measurement Information System scores for anxiety and social isolation, summed severity index of 16 symptoms from the General Symptom Distress Scale, and the total number of symptoms. Self-reported comorbid conditions were measured using the Bayliss tool. General linear models were used to relate symptom measures (one at a time) to age, number of comorbid conditions, level of education, marital status, income meeting needs, and size of metropolitan neighborhood. Additional covariates included site of cancer, its treatment, and whether the cancer was metastatic. Non-Hispanic White survivors ( n = 191) were older and had more comorbid conditions, a higher proportion of metastatic cancers, and higher levels of education and income compared with Hispanic survivors ( n = 168) and non-Hispanic survivors of other races ( n = 41). Compared with the other two groups, Hispanic survivors had the lowest rate of health insurance availability, and non-Hispanic survivors of other races had the lowest social isolation. Age, number of comorbid conditions, and social isolation were significantly associated with number of symptoms, symptom severity, and depression. Age and social isolation were associated with anxiety. In addition, the symptom severity of non-Hispanic White survivors was lower than that of Hispanic survivors and non-Hispanic survivors of other races. These findings highlight the health disparities in symptom burden experienced among cancer survivors when considering their social determinants of health. Assessing these may help clinicians address health disparities in cancer care.

Spatial patterning of Wyoming big sagebrush indicates negative density dependence

AbstractPlants reflect resource use in their spatial patterning. Competition for limited resources—such as available soil water in a dryland ecosystem—drives establishment, growth, and mortality, resulting in shifts of spatial arrangement over time. We characterized the spatial patterning of two big sagebrush (Artemisia tridentata subspecies wyomingensis) communities in the upper Green River Basin of Wyoming, USA. We mapped big sagebrush canopies in two, 100‐m2 sites and calculated plant neighborhoods as the area closer to a target plant than to any other plant. We assumed that neighborhoods were areas in which the target plant dominates resource use. We found that plant neighborhoods had strong, positive correlations with plant size, indicating that larger neighborhoods may access more belowground resources. We also found that the relationships between experienced crowding, that is, crowding index (CI) by an average neighbor, and neighborhood size, were consistently negative regardless of calculation method. We also found that the residuals of a regression of target plant biomass and neighborhood area were strongly related to the CI calculated via all methods. This means that plants with smaller neighborhoods than expected also experience the greatest crowding by an average neighbor. These results are consistent with negative density dependence and show that greater static crowding predicts smaller neighborhoods in two, undisturbed, intermediate‐successional big sagebrush communities. In the future, similar studies of spatial patterning that include interspecific plant–plant interactions will be useful for understanding the relationship between spatial patterning and negative density dependence.

Discriminative Projection Learning With Adaptive Reversed Graph Embedding for Supervised and Semi-Supervised Dimensionality Reduction

The graph embedding based dimensionality reduction (DR) methods have been widely used to overcome the curse of dimensionality problem in high-dimensional data. Especially in pattern recognition and data analysis, they are frequently used to extract features. However, there are two problems in the graph embedding based supervised DR methods: 1) the original data will inevitably be corrupted by noise or outliers, which makes the fixed graph constructed from these data unreliable; and 2) determining how to model the projection for enhancing the discrimination of projected samples. To overcome the above two challenges, this paper proposes a new supervised DR method called discriminative projection learning with adaptive reversed graph embedding (DP-ARGE). Specifically, DP-ARGE learns the graph dynamically in a low-dimensional subspace by using the low-dimensional embedding rather than projected samples for improving its robustness. In addition, a novel discriminative regularization term is designed to enhance the discrimination of the projection samples by using a few pairwise marginal heterogeneous samples rather than the all. To tackle the problem of estimating the neighborhood size parameter of the graph, a simple yet effective automatic parameter estimation strategy based on the density and similarity of data set is proposed. Furthermore, for better handling the data set with a few labels, DP-ARGE is extend to a semi-supervised method (SDP-ARGE) by a neighborhood preserving regularization term. SDP-ARGE can be used to preserve the local manifold structure of the data in the semi-supervised scenario. Comprehensive experiments show the superiority of the proposed methods in several real-world data sets.

The Chinese Lexicon Project II: A megastudy of speeded naming performance for 25,000+ traditional Chinese two-character words.

Using a megastudy approach, (Tse et al.,2017Behavior Research Methods, 49,1503-1519)established a large-scale repository of lexical variables and lexical decision responses for more than 25,000 traditional Chinese two-character words. In the current study, we expand their database by collecting norms for speeded naming reaction times (RTs) and accuracy rates, and compiling more lexical variables (e.g., phonological consistency and semantic neighborhood size). Following Tse et al.'s procedure, about 33 college-aged native Cantonese speakers in Hong Kong read aloud each word. We conducted item-level regression analyses to test the relative predictive power of orthographic variables (e.g., stroke count), phonological variables (e.g., phonological consistency), and semantic variables (e.g., semantic transparency) in naming performance. We also compared the effects of lexical variables on naming performance and Tse et al.'s lexical decision performance to examine the extent to which effects are task-specific or task-general. Freely accessible to the research community, this resource provides a valuable addition to other influential mega-databases, such as the English Lexicon Project (Balota et al., 2004Journal of Experimental Psychology: General, 133, 283-316), and furthers our understanding of Chinese word recognition processes.

A class of doubly stochastic shift operators for random graph signals and their boundedness

A class of doubly stochastic graph shift operators (GSO) is proposed, which is shown to exhibit: (i) lower and upper L2-boundedness for locally stationary random graph signals, (ii) L2-isometry for i.i.d. random graph signals with the asymptotic increase in the incoming neighbourhood size of vertices, and (iii) preservation of the mean of any graph signal – all prerequisites for reliable graph neural networks. These properties are obtained through a statistical consistency analysis of the proposed graph shift operator, and by exploiting the dual role of the doubly stochastic GSO as a Markov (diffusion) matrix and as an unbiased expectation operator. For generality, we consider directed graphs which exhibit asymmetric connectivity matrices. The proposed approach is validated through an example on the estimation of a vector field.

A novel density-based outlier detection method using key attributes

Outlier detection has attracted extensive attention in medical, financial, telecommunications and other fields. Although many related technologies have been proposed, most of them are faced with the problems of the neighborhood size of an object is difficult to determine and the distance in high-dimensional space is unreliable. To overcome these weaknesses, we propose a novel density-based outlier detection method that introduces the concept of Minimum the Sum of Edge Set and other related definitions in key attributes space. Based on the stability of Reverse Minimum the Sum of Edge Set, the proposed method can adaptively select the parameter representing the neighborhood size. In addition, some properties of the proposed local outlier factor are derived. Experiments on synthetic and real-world datasets demonstrate that our method is more effective than the existing outlier detection approaches.

Close Kin Dyads Indicate Intergenerational Dispersal and Barriers.

AbstractThe movement of individuals through continuous space is typically constrained by dispersal ability and dispersal barriers. A range of approaches have been developed to investigate these. Kindisperse is a new approach that infers recent intergenerational dispersal (σ) from close kin dyads and appears particularly useful for investigating taxa that are difficult to observe individually. This study, focusing on the mosquito Aedes aegypti, shows how the same close kin data can also be used for barrier detection. We empirically demonstrate this new extension of the method using genome-wide sequence data from 266 Ae. aegypti. First, we use the spatial distribution of full-sib dyads collected within one generation to infer past movements of ovipositing female mosquitoes. These dyads indicated the relative barrier strengths of two roads and performed favorably against alternative genetic methods for detecting barriers. We then use Kindisperse to quantify recent intergenerational dispersal (-197.1 m generation-1/2) from the difference in variance between the sib and the first cousin spatial distributions and, from this, estimate effective population density (-4,864 km-2). Dispersal estimates showed general agreement with those from mark-release-recapture studies. Barriers, σ, ρ, and neighborhood size (331-526) can inform forthcoming releases of dengue-suppressing Wolbachia bacteria into this mosquito population.

A Feature Extraction Algorithm Based on Optimal Neighborhood Size

The feature extraction of a point cloud fragment model is the basis of fragment splicing, which provides the technical support for research on the segmentation, splicing, and restoration of fragment surfaces. High-quality feature extraction, however, is a complicated process due to the diversity of the surface information of a fragment model. For this subject, a high-efficient point cloud feature extraction method was proposed to address a new method for extracting feature lines. First, the projection distance feature of the point cloud model was calculated to identify the potential feature points. Furthermore, the local information of the possible feature points was used to construct the adaptive neighborhoods for identifying the feature points based on neighborhoods of the model. The clustering fusion of the feature points was proposed according to the discrimination threshold values of the feature points. Finally, the Laplace operator was utilized to refine and connect the feature points to form smooth feature lines. The experimental results showed that the proposed method was automatic, highly efficient, and with good adaptability that could effectively extract the detailed features and construct the complete feature lines. Moreover, results showed that the provided framework could extract the features of simple structure models and be feasible to a certain extent for fragment models with abundant features.

From an agent-based laboratory to the real world: Effects of “neighborhood” size on urban sprawl

Agent-based models (ABMs) have been established as a valuable research tool in the study of complex human-environment systems. However, it is still challenging to produce generalizable and practical results with ABMs for theory development. We use the case of the effects of “neighborhood” size on urban sprawl to illustrate the practice of generating and testing general conclusions from ABMs. In agent-based urban land-change models, homebuyer agents assess land utilities with a bundle of attributes including the quality of focal or zonal neighborhoods. The size of such neighborhoods, or more precisely the spatial assessment units (SAUs), has significant impacts on simulated urban patterns, yet this sensitivity of modeled impacts has not been validated with empirical evidence. Using an ABM that features markets and competitive bidding, we explored how the neighborhood or SAU size affects urban development patterns under various scenarios of homebuyer preferences, which produced a consistent finding that bigger SAUs tended to generate more sprawled patterns. To bring this finding to the real world, we evaluated relationships in data between sprawl indices in US metropolitan areas and the average sizes of their school districts—one type of real-world SAU in neighborhood delineation and a strong input to homebuyers' assessment of housing quality. While the finding has practical policy implications for urban planning and public education, this case exemplifies a path for simple ABMs to support empirically grounded theory development.

Assessing the Impact of Neighborhood Size on Temporal Convolutional Networks for Modeling Land Cover Change

Land cover change (LCC) studies are increasingly using deep learning (DL) modeling techniques. Past studies have leveraged temporal or spatiotemporal sequences of historical LC data to forecast changes with DL models. However, these studies do not adequately assess the association between neighborhood size and DL model capability to forecast LCCs, where neighborhood size refers to the spatial extent captured by each data sample. The objectives of this research study were to: (1) evaluate the effect of neighborhood size on the capacity of DL models to forecast LCCs, specifically Temporal Convolutional Networks (TCN) and Convolutional Neural Networks (CNN-TCN), and (2) assess the effect of auxiliary spatial variables on model capacity to forecast LCCs. First, each model type and neighborhood setting configuration was assessed using data derived from multitemporal MODIS LC for the Regional District of Bulkley-Nechako, Canada, comparing subareas exhibiting different amounts of LCCs with trends obtained for the full region. Next, outcomes were compared with three other study regions. The modeling results were evaluated with three-map comparison measures, where the real-world LC for the next timestep, the real-world LC for the previous timestep, and the forecasted LC for the next year were used to calculate correctly transitioned areas. Across all regions explored, it was observed that increasing neighborhood sizes improved the DL model’s capabilities to forecast short-term LCCs. CNN–TCN models forecasted the most correct LCCs for several regions while reducing error due to quantity when provided additional spatial variables. This study contributes to the systematic exploration of neighborhood sizes on selected spatiotemporal DL techniques for geographic applications.

When move acceptance selection hyper-heuristics outperform Metropolis and elitist evolutionary algorithms and when not

Selection hyper-heuristics (HHs) are automated algorithm selection methodologies that choose between different heuristics during the optimisation process. Recently, selection HHs choosing between a collection of elitist randomised local search heuristics with different neighbourhood sizes have been shown to optimise standard unimodal benchmark functions from evolutionary computation in the optimal expected runtime achievable with the available low-level heuristics. In this paper, we extend our understanding of the performance of HHs to the domain of multimodal optimisation by considering a Move Acceptance HH (MAHH) from the literature that can switch between elitist and non-elitist heuristics during the run. In essence, MAHH is a non-elitist search heuristic that differs from other search heuristics in the source of non-elitism.We first identify the range of parameters that allow MAHH to hillclimb efficiently and prove that it can optimise the standard hillclimbing benchmark function OneMax in the best expected asymptotic time achievable by unbiased mutation-based randomised search heuristics. Afterwards, we use standard multimodal benchmark functions to highlight function characteristics where MAHH outperforms elitist evolutionary algorithms and the well-known Metropolis non-elitist algorithm by quickly escaping local optima, and ones where it does not. Since MAHH is essentially a non-elitist random local search heuristic, the paper is of independent interest to researchers in the fields of artificial intelligence and randomised search heuristics.

Disambiguating Effects of Syllable Position and Neighborhood Size: Contributions of Hanja During Sino-Korean Processing.

A language's writing system offers a complex lens through which to explore its lexicon. Korean's bi-scriptal lexicon comprising its native script Hangul and Chinese Hanja, enables a unique window into what is and is not permissible in the language, as well as a chance to investigate how properties of the written form are reflected in the mental representation of the language. Through a novel priming paradigm, we investigated the effects of Hanja on visual word recognition in Hangul. In particular, we examined the effects of neighborhood size and syllable position of individual Sino-Korean morphemes. Although Hangul is the primary script, literate native Korean speakers are sensitive to the effects of Hanja; they are confronted with a lexicon written primarily in one (Hangul), but deeply influenced by another that is much less visible and commonly used (Hanja). We show that the contributions of Hanja are simply part and parcel of Sino-Korean lexical processing despite the asymmetry in use.

Parzen Window Approximation on Riemannian Manifold

In graph motivated learning, label propagation largely depends on data affinity represented as edges between connected data points. The affinity assignment implicitly assumes even distribution of data on the manifold. This assumption may not hold and may lead to inaccurate metric assignment due to drift towards high-density regions. The drift affected heat kernel based affinity with a globally fixed Parzen window either discards genuine neighbors or forces distant data points to become a member of the neighborhood. This yields a biased affinity matrix. In this paper, the bias due to uneven data sampling on the Riemannian manifold is catered to by a variable Parzen window determined as a function of neighborhood size, ambient dimension, flatness range, etc. Additionally, affinity adjustment is used which offsets the effect of uneven sampling responsible for the bias. An affinity metric which takes into consideration the irregular sampling effect to yield accurate label propagation is proposed. Extensive experiments on synthetic and real-world data sets confirm that the proposed method increases the classification accuracy significantly and outperforms existing Parzen window estimators in graph Laplacian manifold regularization methods.

An enhanced decomposition-based multi-objective evolutionary algorithm with a self-organizing collaborative scheme

The multi-objective evolutionary algorithm based on decomposition (MOEA/D) decomposes a multi-objective optimization problem (MOP) into multiple single-objective subproblems using an aggregation function and optimizes them together using a collaborative approach. MOEA/D exhibits extraordinary optimization ability in solving MOPs. However, the algorithm’s performance is hampered by the single evolutionary operator, fixed control parameters, and fixed neighborhood architecture used in primal MOEA/D. To obtain the balance between convergence and diversity, a self-organizing collaborative scheme (SCS) is designed to achieve the best combination of differential evolution (DE) operator, control parameters, and neighborhood size in this paper. The SCS is integrated into MOEA/D to form the MOEA/D-SCS. In MOEA/D-SCS, firstly, the evolutionary process is separated into three stages: early, middle, and late, and different indicators are utilized to screen elite and inferior solutions for the early and late stages. Secondly, different evolutionary strategies are developed for the three stages and the various types of solutions to help boost the algorithm’s efficiency. Then, to enhance convergence and diversity in different evolutionary stages, a neighborhood reconstruction strategy is proposed to assign different neighborhood sizes according to the neighborhood information. Finally, the capability of MOEA/D-SCS was tested using three standard test suites, ZDT, DTLZ, and UF, and a MOP of sewage treatment. The simulation results demonstrate that the proposed MOEA/D-SCS can achieve better optimization results and show stronger optimization ability in most cases compared to state-of-the-art evolutionary algorithms, such as MOEA/D-DE, MOEA/D-STM, MOEA/D-FRRMAB, MOEA/D-DRA, NSGA-II, and dMOPSO, especially in problems with multiple peaks, complex Pareto fronts, and practical engineering optimization problems.

How does the structure of data impact cell-cell similarity? Evaluating how structural properties influence the performance of proximity metrics in single cell RNA-seq data.

Accurately identifying cell-populations is paramount to the quality of downstream analyses and overall interpretations of single-cell RNA-seq (scRNA-seq) datasets but remains a challenge. The quality of single-cell clustering depends on the proximity metric used to generate cell-to-cell distances. Accordingly, proximity metrics have been benchmarked for scRNA-seq clustering, typically with results averaged across datasets to identify a highest performing metric. However, the 'best-performing' metric varies between studies, with the performance differing significantly between datasets. This suggests that the unique structural properties of an scRNA-seq dataset, specific to the biological system under study, have a substantial impact on proximity metric performance. Previous benchmarking studies have omitted to factor the structural properties into their evaluations. To address this gap, we developed a framework for the in-depth evaluation of the performance of 17 proximity metrics with respect to core structural properties of scRNA-seq data, including sparsity, dimensionality, cell-population distribution and rarity. We find that clustering performance can be improved substantially by the selection of an appropriate proximity metric and neighbourhood size for the structural properties of a dataset, in addition to performing suitable pre-processing and dimensionality reduction. Furthermore, popular metrics such as Euclidean and Manhattan distance performed poorly in comparison to several lessor applied metrics, suggesting that the default metric for many scRNA-seq methods should be re-evaluated. Our findings highlight the critical nature of tailoring scRNA-seq analyses pipelines to the dataset under study and provide practical guidance for researchers looking to optimize cell-similarity search for the structural properties of their own data.

Prevalence norms for 40,777 Catalan words: An online megastudy of vocabulary size

In this study, we present word prevalence data (i.e., the number of people who know a given word) for 40,777 Catalan words. An online massive visual lexical decision task involving more than 200,000 native speakers of this language was carried out. The characteristics of the participants as well as those of the words which mostly influence word knowledge were examined. Regarding the participants, the analysis of the data revealed that their age was the main factor influencing vocabulary size, followed by their educational level and other variables such as the number of languages spoken and their level of proficiency in Catalan. Concerning the words, by far the most determining factor was lexical frequency, with a minor influence of both length and the size of the orthographic neighborhood. These data mainly agree with those reported in other languages in which the same variables have been analyzed (Dutch, English, and Spanish, thus far). Therefore, the list is increased with Catalan, a language which, due to its use in an essentially bilingual context, is of special interest to researchers interested in the field of bilingualism and second language acquisition.

Remote sensing image classification based on object-oriented convolutional neural network

Remote sensing image classification is of great importance for urban development and planning. The need for higher classification accuracy has led to improvements in classification technology. In this research, Landsat 8 images are used as experimental data, and Wuhan, Chengde and Tongchuan are selected as research areas. The best neighborhood window size of the image patch and band combination method are selected based on two sets of comparison experiments. Then, an object-oriented convolutional neural network (OCNN) is used as a classifier. The experimental results show that the classification accuracy of the OCNN classifier is 6% higher than that of an SVM classifier and 5% higher than that of a convolutional neural network classifier. The graph of the classification results of the OCNN is more continuous than the plots obtained with the other two classifiers, and there are few fragmentations observed for most of the category. The OCNN successfully solves the salt and pepper problem and improves the classification accuracy to some extent, which verifies the effectiveness of the proposed object-oriented model.

Effects of neighborhood-based collaborative filtering parameters on their blockbuster bias performances

Collaborative filtering algorithms are efficient tools for providing recommendations with reasonable accuracy performances to individuals. However, the previous research has realized that these algorithms are undesirably biased towards blockbuster items. i.e., both popular and highly-liked items, in their recommendations, resulting in recommendation lists dominated by such blockbuster items. As one most prominent types of collaborative filtering approaches, neighborhood-based algorithms aim to produce recommendations based on neighborhoods constructed based on similarities between users or items. Therefore, the utilized similarity function and the size of the neighborhoods are critical parameters on their recommendation performances. This study considers three well-known similarity functions, i.e., Pearson, Cosine, and Mean Squared Difference, and varying neighborhood sizes and observes how they affect the algorithms’ blockbuster bias and accuracy performances. The extensive experiments conducted on two benchmark data collections conclude that as the size of neighborhoods decreases, these algorithms generally become more vulnerable to blockbuster bias while their accuracy increases. The experimental works also show that using the Cosine metric is superior to other similarity functions in producing recommendations where blockbuster bias is treated more; however, it leads to having unqualified recommendations in terms of predictive accuracy as they are usually conflicting goals.

A Simultaneous Multiparametric 18F-FDG PET/MRI Radiomics Model for the Diagnosis of Triple Negative Breast Cancer.

Simple SummaryIn this study, we aimed to build a machine-learning predictive model for the identification of triple negative breast cancer, the most aggressive subtype, using quantitative parameters and radiomics features extracted from tumor lesions on hybrid PET/MRI. The good performance of the model supports the hypothesis that hybrid PET/MRI can provide quantitative data able to non-invasively detect tumor biological characteristics using artificial intelligence software and further encourages the conduction of additional studies for this purpose.Purpose: To investigate whether a machine learning (ML)-based radiomics model applied to 18F-FDG PET/MRI is effective in molecular subtyping of breast cancer (BC) and specifically in discriminating triple negative (TN) from other molecular subtypes of BC. Methods: Eighty-six patients with 98 BC lesions (Luminal A = 10, Luminal B = 51, HER2+ = 12, TN = 25) were included and underwent simultaneous 18F-FDG PET/MRI of the breast. A 3D segmentation of BC lesion was performed on T2w, DCE, DWI and PET images. Quantitative diffusion and metabolic parameters were calculated and radiomics features extracted. Data were selected using the LASSO regression and used by a fine gaussian support vector machine (SVM) classifier with a 5-fold cross validation for identification of TNBC lesions. Results: Eight radiomics models were built based on different combinations of quantitative parameters and/or radiomic features. The best performance (AUROC 0.887, accuracy 82.8%, sensitivity 79.7%, specificity 86%, PPV 85.3%, NPV 80.8%) was found for the model combining first order, neighborhood gray level dependence matrix and size zone matrix-based radiomics features extracted from ADC and PET images. Conclusion: A ML-based radiomics model applied to 18F-FDG PET/MRI is able to non-invasively discriminate TNBC lesions from other BC molecular subtypes with high accuracy. In a future perspective, a “virtual biopsy” might be performed with radiomics signatures.