Genotype-phenotype (G-P) analyses for complex morphological traits typically utilize simple, predetermined anatomical measures or features derived via unsupervised dimension reduction techniques (e.g. principal component analysis (PCA) or eigen-shapes). Despite the popularity of these approaches, they do not necessarily reveal axes of phenotypic variation that are genetically relevant. Therefore, we introduce a framework to optimize phenotyping for G-P analyses, such as genome-wide association studies (GWAS) of common variants or rare variant association studies (RVAS) of rare variants. Our strategy is two-fold: (i) we construct a multidimensional feature space spanning a wide range of phenotypic variation, and (ii) within this feature space, we use an optimization algorithm to search for directions or feature combinations that are genetically enriched. To test our approach, we examine human facial shape in the context of GWAS and RVAS. In GWAS, we optimize for phenotypes exhibiting high heritability, estimated from either family data or genomic relatedness measured in unrelated individuals. In RVAS, we optimize for the skewness of phenotype distributions, aiming to detect commingled distributions that suggest single or few genomic loci with major effects. We compare our approach with eigen-shapes as baseline in GWAS involving 8246 individuals of European ancestry and in gene-based tests of rare variants with a subset of 1906 individuals. After applying linkage disequilibrium score regression to our GWAS results, heritability-enriched phenotypes yielded the highest SNP heritability, followed by eigen-shapes, while commingling-based traits displayed the lowest SNP heritability. Heritability-enriched phenotypes also exhibited higher discovery rates, identifying the same number of independent genomic loci as eigen-shapes with a smaller effective number of traits. For RVAS, commingling-based traits resulted in more genes passing the exome-wide significance threshold than eigen-shapes, while heritability-enriched phenotypes lead to only a few associations. Overall, our results demonstrate that optimized phenotyping allows for the extraction of genetically relevant traits that can specifically enhance discovery efforts of common and rare variants, as evidenced by their increased power in facial GWAS and RVAS.

MOTIVATION: Combining omics and images can lead to a more comprehensive clustering of individuals than classic single-view approaches. Among the various approaches for multi-view clustering, nonnegative matrix tri-factorization (NMTF) and nonnegative Tucker decomposition (NTD) are advantageous in learning low-rank embeddings with promising interpretability. Besides, there is a need to handle unwanted drivers of clusterings (i.e. confounders). RESULTS: In this work, we introduce a novel multi-view clustering method based on NMTF and NTD, named INMTD, which integrates omics and 3D imaging data to derive unconfounded subgroups of individuals. According to the adjusted Rand index, INMTD outperformed other clustering methods on a synthetic dataset with known clusters. In the application to real-life facial-genomic data, INMTD generated biologically relevant embeddings for individuals, genetics, and facial morphology. By removing confounded embedding vectors, we derived an unconfounded clustering with better internal and external quality; the genetic and facial annotations of each derived subgroup highlighted distinctive characteristics. In conclusion, INMTD can effectively integrate omics data and 3D images for unconfounded clustering with biologically meaningful interpretation. AVAILABILITY AND IMPLEMENTATION: INMTD is freely available at https://github.com/ZuqiLi/INMTD.

Unsupervised learning, particularly clustering, plays a pivotal role in disease subtyping and patient stratification, especially with the abundance of large-scale multi-omics data. Deep learning models, such as variational autoencoders (VAEs), can enhance clustering algorithms by leveraging inter-individual heterogeneity. However, the impact of confounders-external factors unrelated to the condition, e.g. batch effect or age-on clustering is often overlooked, introducing bias and spurious biological conclusions. In this work, we introduce four novel VAE-based deconfounding frameworks tailored for clustering multi-omics data. These frameworks effectively mitigate confounding effects while preserving genuine biological patterns. The deconfounding strategies employed include (i) removal of latent features correlated with confounders, (ii) a conditional VAE, (iii) adversarial training, and (iv) adding a regularization term to the loss function. Using real-life multi-omics data from The Cancer Genome Atlas, we simulated various confounding effects (linear, nonlinear, categorical, mixed) and assessed model performance across 50 repetitions based on reconstruction error, clustering stability, and deconfounding efficacy. Our results demonstrate that our novel models, particularly the conditional multi-omics VAE (cXVAE), successfully handle simulated confounding effects and recover biologically driven clustering structures. cXVAE accurately identifies patient labels and unveils meaningful pathological associations among cancer types, validating deconfounded representations. Furthermore, our study suggests that some of the proposed strategies, such as adversarial training, prove insufficient in confounder removal. In summary, our study contributes by proposing innovative frameworks for simultaneous multi-omics data integration, dimensionality reduction, and deconfounding in clustering. Benchmarking on open-access data offers guidance to end-users, facilitating meaningful patient stratification for optimized precision medicine.

Human facial shape, while strongly heritable, involves both genetic and structural complexity, necessitating precise phenotyping for accurate assessment. Common phenotyping strategies include simplifying 3D facial features into univariate traits such as anthropometric measurements (e.g., inter-landmark distances), unsupervised dimensionality reductions (e.g., principal component analysis (PCA) and auto-encoder (AE) approaches), and assessing resemblance to particular facial gestalts (e.g., syndromic facial archetypes). This study provides a comparative assessment of these strategies in genome-wide association studies (GWASs) of 3D facial shape. Specifically, we investigated inter-landmark distances, PCA and AE-derived latent dimensions, and facial resemblance to random, extreme, and syndromic gestalts within a GWAS of 8,426 individuals of recent European ancestry. Inter-landmark distances exhibit the highest SNP-based heritability as estimated via LD score regression, followed by AE dimensions. Conversely, resemblance scores to extreme and syndromic facial gestalts display the lowest heritability, in line with expectations. Notably, the aggregation of multiple GWASs on facial resemblance to random gestalts reveals the highest number of independent genetic loci. This novel, easy-to-implement phenotyping approach holds significant promise for capturing genetically relevant morphological traits derived from complex biomedical imaging datasets, and its applications extend beyond faces. Nevertheless, these different phenotyping strategies capture different genetic influences on craniofacial shape. Thus, it remains valuable to explore these strategies individually and in combination to gain a more comprehensive understanding of the genetic factors underlying craniofacial shape and related traits.

A genome-wide association study (GWAS) of a complex, multi-dimensional morphological trait, such as the human face, typically relies on predefined and simplified phenotypic measurements, such as inter-landmark distances and angles. These measures are predominantly designed by human experts based on perceived biological or clinical knowledge. To avoid use handcrafted phenotypes (i.e., a priori expert-identified phenotypes), alternative automatically extracted phenotypic descriptors, such as features derived from dimension reduction techniques (e.g., principal component analysis), are employed. While the features generated by such computational algorithms capture the geometric variations of the biological shape, they are not necessarily genetically relevant. Therefore, genetically informed data-driven phenotyping is desirable. Here, we propose an approach where phenotyping is done through a data-driven optimization of trait heritability, defined as the degree of variation in a phenotypic trait in a population that is due to genetic variation. The resulting phenotyping process consists of two steps: 1) constructing a feature space that models shape variations using dimension reduction techniques, and 2) searching for directions in the feature space exhibiting high trait heritability using a genetic search algorithm (i.e., heuristic inspired by natural selection). We show that the phenotypes resulting from the proposed trait heritability-optimized training differ from those of principal components in the following aspects: 1) higher trait heritability, 2) higher SNP heritability, and 3) identification of the same number of independent genetic loci with a smaller number of effective traits. Our results demonstrate that data-driven trait heritability-based optimization enables the automatic extraction of genetically relevant phenotypes, as shown by their increased power in genome-wide association scans.

Face recognition is a widely accepted biometric verification tool, as the face contains a lot of information about the identity of a person. In this study, a 2-step neural-based pipeline is presented for matching 3D facial shape to multiple DNA-related properties (sex, age, BMI and genomic background). The first step consists of a triplet loss-based metric learner that compresses facial shape into a lower dimensional embedding while preserving information about the property of interest. Most studies in the field of metric learning have only focused on 2D Euclidean data. In this work, geometric deep learning is employed to learn directly from 3D facial meshes. To this end, spiral convolutions are used along with a novel mesh-sampling scheme that retains uniformly sampled 3D points at different levels of resolution. The second step is a multi-biometric fusion by a fully connected neural network. The network takes an ensemble of embeddings and property labels as input and returns genuine and imposter scores. Since embeddings are accepted as an input, there is no need to train classifiers for the different properties and available data can be used more efficiently. Results obtained by a to-fold cross-validation for biometric verification show that combining multiple properties leads to stronger biometric systems. Furthermore, the proposed neural-based pipeline outperforms a linear baseline, which consists of principal component analysis, followed by classification with linear support vector machines and a Naïve Bayes-based score-fuser.

Statistical shape modeling is a powerful tool for visualizing and quantifying geometric and functional patterns of the heart. After myocardial infarction (MI), the left ventricle typically remodels in response to physiological challenges. Several methods have been proposed in the literature to describe statistical shape changes. Which method best characterizes left ventricular remodeling after MI is an open research question. A better descriptor of remodeling is expected to provide a more accurate evaluation of disease status in MI patients. We therefore designed a challenge to test shape characterization in MI given a set of three-dimensional left ventricular surface points. The training set comprised 100 MI patients, and 100 asymptomatic volunteers (AV). The challenge was initiated in 2015 at the Statistical Atlases and Computational Models of the Heart workshop, in conjunction with the MICCAI conference. The training set with labels was provided to participants, who were asked to submit the likelihood of MI from a different (validation) set of 200 cases (100 AV and 100 MI). Sensitivity, specificity, accuracy and area under the receiver operating characteristic curve were used as the outcome measures. The goals of this challenge were to (1) establish a common dataset for evaluating statistical shape modeling algorithms in MI, and (2) test whether statistical shape modeling provides additional information characterizing MI patients over standard clinical measures. Eleven groups with a wide variety of classification and feature extraction approaches participated in this challenge. All methods achieved excellent classification results with accuracy ranges from 0.83 to 0.98. The areas under the receiver operating characteristic curves were all above 0.90. Four methods showed significantly higher performance than standard clinical measures. The dataset and software for evaluation are available from the Cardiac Atlas Project website1.

Research in face recognition has continuously been challenged by extrinsic (head pose, lighting conditions) and intrinsic (facial expression, aging) sources of variability. While many survey papers on face recognition exist, in this paper, we focus on a comparative study of 3-D face recognition under expression variations. As a first contribution, 3-D face databases with expressions are listed, and the most important ones are briefly presented and their complexity is quantified using the iterative closest point (ICP) baseline recognition algorithm. This allows to rank the databases according to their inherent difficulty for face-recognition tasks. This analysis reveals that the FRGC v2 database can be considered as the most challenging because of its size, the presence of expressions and outliers, and the time lapse between the recordings. Therefore, we recommend to use this database as a reference database to evaluate (expression-invariant) 3-D face-recognition algorithms. We also determine and quantify the most important factors that influence the performance. It appears that performance decreases 1) with the degree of nonfrontal pose, 2) for certain expression types, 3) with the magnitude of the expressions, 4) with an increasing number of expressions, and 5) for a higher number of gallery subjects. Future 3-D face-recognition algorithms should be evaluated on the basis of all these factors. As the second contribution, a survey of published 3-D face-recognition methods that deal with expression variations is given. These methods are subdivided into three classes depending on the way the expressions are handled. Region-based methods use expression-stable regions only, while other methods model the expressions either using an isometric or a statistical model. Isometric models assume the deformation because of expression variation to be (locally) isometric, meaning that the deformation preserves lengths along the surface. Statistical models learn how the facial soft tissue deforms during expressions based on a training database with expression labels. Algorithmic performances are evaluated by the comparison of recognition rates for identification and verification. No statistical significant differences in class performance are found between any pair of classes.

Since most 3D cameras cannot capture the complete 3D face, an important challenge in 3D face recognition is the comparison of two 3D facial surfaces with little or no over lap. In this paper, a local feature method is presented to tackle this challenge exploiting the symmetry of the human face. Features are located and described using an extension of SIFT for meshes (meshSIFT). As such, features are localized as extrema in the curvature scale space of the in put mesh, and are described by concatenating histograms of shape indices and slant angles of the neighborhood. For 3D face scans with sufficient overlap, the number of matching meshSIFT features is a reliable measure for face recognition purposes. However, as the feature descriptor is not symmetrical, features on one face are not matched with their symmetrical counterpart on another face impeding their feasibility for comparison efface scans with limited or no (left-right) overlap. In order to alleviate this problem, facial symmetry could be used to increase the overlap between two face scans by mirroring one of both faces w.r.t. an arbitrary plane. As this would increase the computational demand, this paper proposes an efficient approach to de scribe the features of a mirrored face by mirroring the mesh SIFT descriptors of the input face. The presented method is validated on the data of the "SHREC '11: Face Scans" contest, containing many partial scans. This resulted in a recognition rate of 98.6% and a mean average precision of 93.3%, clearly outperforming all other participants in the challenge.