Capturing phenotypes for precision medicine

Hybrid natural language processing tool for semantic annotation of medical texts in Spanish

BACKGROUND

Natural language processing (NLP) enables the extraction of information embedded within unstructured texts, such as clinical case reports and trial eligibility criteria. By identifying relevant medical concepts, NLP facilitates the generation of structured and actionable data, supporting complex tasks like cohort identification and the analysis of clinical records. To accomplish those tasks, we introduce a deep learning-based and lexicon-based named entity recognition (NER) tool for texts in Spanish. It performs medical NER and normalization, medication information extraction and detection of temporal entities, negation and speculation, and temporality or experiencer attributes (Age, Contraindicated, Negated, Speculated, Hypothetical, Future, Family_member, Patient and Other). We built the tool with a dedicated lexicon and rules adapted from NegEx and HeidelTime. Using these resources, we annotated a corpus of 1200 texts, with high inter-annotator agreement (average F1 = 0.841% ± 0.045 for entities, and average F1 = 0.881% ± 0.032 for attributes). We used this corpus to train Transformer-based models (RoBERTa-based models, mBERT and mDeBERTa). We integrated them with the dictionary-based system in a hybrid tool, and distribute the models via the Hugging Face hub. For an internal validation, we used a held-out test set and conducted an error analysis. For an external validation, eight medical professionals evaluated the system by revising the annotation of 200 new texts not used in development.

RESULTS

In the internal validation, the models yielded F1 values up to 0.915. In the external validation with 100 clinical trials, the tool achieved an average F1 score of 0.858 (± 0.032); and in 100 anonymized clinical cases, it achieved an average F1 score of 0.910 (± 0.019).

CONCLUSIONS

The tool is available at https://claramed.csic.es/medspaner . We also release the code ( https://github.com/lcampillos/medspaner ) and the annotated corpus to train the models.

Unsupervised clustering approach to assess heterogeneity of treatment effects across patient phenotypes in randomized clinical trials

BACKGROUND

Primary results from randomized clinical trials (RCT) only inform on the average treatment effect in the studied population, and it is critical to understand how treatment effect varies across subpopulations. In this paper we describe a clustering-based approach for the assessment of Heterogeneity of Treatment Effect (HTE) over patient phenotypes, which maintains the unsupervised nature of classical subgroup analysis while jointly accounting for relevant patient characteristics.

METHODS

We applied phenotype-based stratification in the ENGAGE AF-TIMI 48 trial, a non-inferiority trial comparing the effects of higher-dose edoxaban regimen (direct anticoagulant) versus warfarin (vitamin K antagonist) on a composite endpoint of stroke and systemic embolism in 14,062 patients with atrial fibrillation.

RESULTS

We identified three distinct phenotypes: non-white participants, mostly from Asia (A); white participants without previous use of vitamin-K antagonists (B); and white participants with previous use of vitamin-K antagonist (C). The effect of the higher-dose edoxaban regimen vs warfarin significantly varied over phenotypes (p for interaction = 0.03) with the strongest benefit in cluster A (HR = 0.72, 95 % CI: 0.52-1.00), moderate effect in cluster B (HR = 0.80, 95 % CI: 0.61, 1.06) and no observed effect in cluster C (HR = 1.01, 95 % CI: 0.80, 1.27).

CONCLUSIONS

Assessing HTE over patients' phenotypes might represent a relevant complement to other stratification approaches to elucidate results from subgroups analyses, especially in those settings where an overwhelming superiority overall effect was not observed. Cluster analysis allows a clear discrimination of patients with direct interpretability of who are the patients that would most benefit from the investigated strategy or treatment.

An interpretable cluster-based logistic regression model, with application to the characterization of response to therapy in severe eosinophilic asthma

Asthma is a disease characterized by chronic airway hyperresponsiveness and inflammation, with signs of variable airflow limitation and impaired lung function leading to respiratory symptoms such as shortness of breath, chest tightness and cough. Eosinophilic asthma is a distinct phenotype that affects more than half of patients diagnosed with severe asthma. It can be effectively treated with monoclonal antibodies targeting specific immunological signaling pathways that fuel the inflammation underlying the disease, particularly Interleukin-5 (IL-5), a cytokine that plays a crucial role in asthma. In this study, we propose a data analysis pipeline aimed at identifying subphenotypes of severe eosinophilic asthma in relation to response to therapy at follow-up, which could have great potential for use in routine clinical practice. Once an optimal partition of patients into subphenotypes has been determined, the labels indicating the group to which each patient has been assigned are used in a novel way. For each input variable in a specialized logistic regression model, a clusterwise effect on response to therapy is determined by an appropriate interaction term between the input variable under consideration and the cluster label. We show that the clusterwise odds ratios can be meaningfully interpreted conditional on the cluster label. In this way, we can define an effect measure for the response variable for each input variable in each of the groups identified by the clustering algorithm, which is not possible in standard logistic regression because the effect of the reference class is aliased with the overall intercept. The interpretability of the model is enforced by promoting sparsity, a goal achieved by learning interactions in a hierarchical manner using a special group-Lasso technique. In addition, valid expressions are provided for computing odds ratios in the unusual parameterization used by the sparsity-promoting algorithm. We show how to apply the proposed data analysis pipeline to the problem of sub-phenotyping asthma patients also in terms of quality of response to therapy with monoclonal antibodies.

Integrating deep phenotyping with genetic analysis: a comprehensive workflow for diagnosis and management of rare bone diseases

Phenotypes play a fundamental role in medical genetics, serving as external manifestations of underlying genotypes. Deep phenotyping, a cornerstone of precision medicine, involves precise multi-system phenotype assessments, facilitating disease subtyping and genetic understanding. Despite their significance, the field lacks standardized protocols for accurate phenotype evaluation, hindering clinical comprehension and research comparability. We present a comprehensive workflow of deep phenotyping for rare bone diseases from the Genetics Clinic of Skeletal Deformity at Peking Union Medical College Hospital. Our workflow integrates referral, informed consent, and detailed phenotype evaluation through HPO standards, capturing nuanced phenotypic characteristics using clinical examinations, questionnaires, and multimedia documentation. Genetic testing and counseling follow, based on deep phenotyping results, ensuring personalized interventions. Multidisciplinary team consultations facilitate comprehensive patient care and clinical guideline development. Regular follow-up visits emphasize dynamic phenotype reassessment, ensuring treatment strategies remain responsive to evolving patient needs. In conclusion, this study highlights the importance of deep phenotyping in rare bone diseases, offering a standardized framework for phenotype evaluation, genetic analysis, and multidisciplinary intervention. By enhancing clinical care and research outcomes, this approach contributes to the advancement of precision medicine in the field of medical genetics.

Creating High-Quality Synthetic Health Data: Framework for Model Development and Validation

BACKGROUND

Electronic health records are a valuable source of patient information that must be properly deidentified before being shared with researchers. This process requires expertise and time. In addition, synthetic data have considerably reduced the restrictions on the use and sharing of real data, allowing researchers to access it more rapidly with far fewer privacy constraints. Therefore, there has been a growing interest in establishing a method to generate synthetic data that protects patients' privacy while properly reflecting the data.

OBJECTIVE

This study aims to develop and validate a model that generates valuable synthetic longitudinal health data while protecting the privacy of the patients whose data are collected.

METHODS

We investigated the best model for generating synthetic health data, with a focus on longitudinal observations. We developed a generative model that relies on the generalized canonical polyadic (GCP) tensor decomposition. This model also involves sampling from a latent factor matrix of GCP decomposition, which contains patient factors, using sequential decision trees, copula, and Hamiltonian Monte Carlo methods. We applied the proposed model to samples from the MIMIC-III (version 1.4) data set. Numerous analyses and experiments were conducted with different data structures and scenarios. We assessed the similarity between our synthetic data and the real data by conducting utility assessments. These assessments evaluate the structure and general patterns present in the data, such as dependency structure, descriptive statistics, and marginal distributions. Regarding privacy disclosure, our model preserves privacy by preventing the direct sharing of patient information and eliminating the one-to-one link between the observed and model tensor records. This was achieved by simulating and modeling a latent factor matrix of GCP decomposition associated with patients.

RESULTS

The findings show that our model is a promising method for generating synthetic longitudinal health data that is similar enough to real data. It can preserve the utility and privacy of the original data while also handling various data structures and scenarios. In certain experiments, all simulation methods used in the model produced the same high level of performance. Our model is also capable of addressing the challenge of sampling patients from electronic health records. This means that we can simulate a variety of patients in the synthetic data set, which may differ in number from the patients in the original data.

CONCLUSIONS

We have presented a generative model for producing synthetic longitudinal health data. The model is formulated by applying the GCP tensor decomposition. We have provided 3 approaches for the synthesis and simulation of a latent factor matrix following the process of factorization. In brief, we have reduced the challenge of synthesizing massive longitudinal health data to synthesizing a nonlongitudinal and significantly smaller data set.

Analyzing Biomedical Datasets with Symbolic Tree Adaptive Resonance Theory

Biomedical datasets distill many mechanisms of human diseases, linking diseases to genes and phenotypes (signs and symptoms of disease), genetic mutations to altered protein structures, and altered proteins to changes in molecular functions and biological processes. It is desirable to gain new insights from these data, especially with regard to the uncovering of hierarchical structures relating disease variants. However, analysis to this end has proven difficult due to the complexity of the connections between multi-categorical symbolic data. This article proposes symbolic tree adaptive resonance theory (START), with additional supervised, dual-vigilance (DV-START), and distributed dual-vigilance (DDV-START) formulations, for the clustering of multi-categorical symbolic data from biomedical datasets by demonstrating its utility in clustering variants of Charcot–Marie–Tooth disease using genomic, phenotypic, and proteomic data.

Dissecting Schizotypy and Its Association With Cognition and Polygenic Risk for Schizophrenia in a Nonclinical Sample

Schizotypy is a multidimensional construct that captures a continuum of risk for developing schizophrenia-spectrum psychopathology. Existing 3-factor models of schizotypy, consisting of positive, negative, and disorganized dimensions have yielded mixed evidence of genetic continuity with schizophrenia using polygenic risk scores. Here, we propose an approach that involves splitting positive and negative schizotypy into more specific subdimensions that are phenotypically continuous with distinct positive symptoms and negative symptoms recognized in clinical schizophrenia. We used item response theory to derive high-precision estimates of psychometric schizotypy using 251 self-report items obtained from a non-clinical sample of 727 (424 females) adults. These subdimensions were organized hierarchically using structural equation modeling into 3 empirically independent higher-order dimensions enabling associations with polygenic risk for schizophrenia to be examined at different levels of phenotypic generality and specificity. Results revealed that polygenic risk for schizophrenia was associated with variance specific to delusional experiences (γ = 0.093, P = .001) and reduced social interest and engagement (γ = 0.076, P = .020), and these effects were not mediated via the higher-order general, positive, or negative schizotypy factors. We further fractionated general intellectual functioning into fluid and crystallized intelligence in 446 (246 females) participants that underwent onsite cognitive assessment. Polygenic risk scores explained 3.6% of the variance in crystallized intelligence. Our precision phenotyping approach could be used to enhance the etiologic signal in future genetic association studies and improve the detection and prevention of schizophrenia-spectrum psychopathology.

Anatomy and the type concept in biology show that ontologies must be adapted to the diagnostic needs of research

BACKGROUND

In times of exponential data growth in the life sciences, machine-supported approaches are becoming increasingly important and with them the need for FAIR (Findable, Accessible, Interoperable, Reusable) and eScience-compliant data and metadata standards. Ontologies, with their queryable knowledge resources, play an essential role in providing these standards. Unfortunately, biomedical ontologies only provide ontological definitions that answer What is it? questions, but no method-dependent empirical recognition criteria that answer How does it look?

QUESTIONS

Consequently, biomedical ontologies contain knowledge of the underlying ontological nature of structural kinds, but often lack sufficient diagnostic knowledge to unambiguously determine the reference of a term.

RESULTS

We argue that this is because ontology terms are usually textually defined and conceived as essentialistic classes, while recognition criteria often require perception-based definitions because perception-based contents more efficiently document and communicate spatial and temporal information-a picture is worth a thousand words. Therefore, diagnostic knowledge often must be conceived as cluster classes or fuzzy sets. Using several examples from anatomy, we point out the importance of diagnostic knowledge in anatomical research and discuss the role of cluster classes and fuzzy sets as concepts of grouping needed in anatomy ontologies in addition to essentialistic classes. In this context, we evaluate the role of the biological type concept and discuss its function as a general container concept for groupings not covered by the essentialistic class concept.

CONCLUSIONS

We conclude that many recognition criteria can be conceptualized as text-based cluster classes that use terms that are in turn based on perception-based fuzzy set concepts. Finally, we point out that only if biomedical ontologies model also relevant diagnostic knowledge in addition to ontological knowledge, they will fully realize their potential and contribute even more substantially to the establishment of FAIR and eScience-compliant data and metadata standards in the life sciences.

PhenoRerank: A re-ranking model for phenotypic concept recognition pre-trained on human phenotype ontology

The study aims at developing a neural network model to improve the performance of Human Phenotype Ontology (HPO) concept recognition tools. We used the terms, definitions, and comments about the phenotypic concepts in the HPO database to train our model. The document to be analyzed is first split into sentences and annotated with a base method to generate candidate concepts. The sentences, along with the candidate concepts, are then fed into the pre-trained model for re-ranking. Our model comprises the pre-trained BlueBERT and a feature selection module, followed by a contrastive loss. We re-ranked the results generated by three robust HPO annotation tools and compared the performance against most of the existing approaches. The experimental results show that our model can improve the performance of the existing methods. Significantly, it boosted 3.0% and 5.6% in F1 score on the two evaluated datasets compared with the base methods. It removed more than 80% of the false positives predicted by the base methods, resulting in up to 18% improvement in precision. Our model utilizes the descriptive data in the ontology and the contextual information in the sentences for re-ranking. The results indicate that the additional information and the re-ranking model can significantly enhance the precision of HPO concept recognition compared with the base method.

Characterizing Long COVID: Deep Phenotype of a Complex Condition

BACKGROUND

Numerous publications describe the clinical manifestations of post-acute sequelae of SARS-CoV-2 (PASC or "long COVID"), but they are difficult to integrate because of heterogeneous methods and the lack of a standard for denoting the many phenotypic manifestations. Patient-led studies are of particular importance for understanding the natural history of COVID-19, but integration is hampered because they often use different terms to describe the same symptom or condition. This significant disparity in patient versus clinical characterization motivated the proposed ontological approach to specifying manifestations, which will improve capture and integration of future long COVID studies.

METHODS

The Human Phenotype Ontology (HPO) is a widely used standard for exchange and analysis of phenotypic abnormalities in human disease but has not yet been applied to the analysis of COVID-19.

FUNDING

We identified 303 articles published before April 29, 2021, curated 59 relevant manuscripts that described clinical manifestations in 81 cohorts three weeks or more following acute COVID-19, and mapped 287 unique clinical findings to HPO terms. We present layperson synonyms and definitions that can be used to link patient self-report questionnaires to standard medical terminology. Long COVID clinical manifestations are not assessed consistently across studies, and most manifestations have been reported with a wide range of synonyms by different authors. Across at least 10 cohorts, authors reported 31 unique clinical features corresponding to HPO terms; the most commonly reported feature was Fatigue (median 45.1%) and the least commonly reported was Nausea (median 3.9%), but the reported percentages varied widely between studies.

INTERPRETATION

Translating long COVID manifestations into computable HPO terms will improve analysis, data capture, and classification of long COVID patients. If researchers, clinicians, and patients share a common language, then studies can be compared/pooled more effectively. Furthermore, mapping lay terminology to HPO will help patients assist clinicians and researchers in creating phenotypic characterizations that are computationally accessible, thereby improving the stratification, diagnosis, and treatment of long COVID.

FUNDING

U24TR002306; UL1TR001439; P30AG024832; GBMF4552; R01HG010067; UL1TR002535; K23HL128909; UL1TR002389; K99GM145411.

Systematic Review: Monoclonal Antibody-Induced Subacute Cutaneous Lupus Erythematosus

Background

Subacute cutaneous lupus erythematosus (SCLE) lacks consensus diagnostic criteria and the pathogenesis is poorly understood. There are increasing reports of SCLE induced by monoclonal antibodies (mAbs), but there are limited data on the aetiology, clinical characteristics and natural course of this disease.

Methods

We devised a set of diagnostic criteria for SCLE in collaboration with a multinational, multispecialty panel. This systematic review employed a two-layered search strategy of five databases for cases of mAb-induced SCLE (PROSPERO registered protocol CRD42019116521). To explore the relationship between relative mAb use and the number of SCLE cases reported, the estimated number of mAb users was modelled from 2013 to 2018 global commercial data and estimated annual therapy costs.

Results

From 40 papers, we identified 52 cases of mAb-induced SCLE, occurring in a cohort that was 73% female and with a median age of 61 years. Fifty percent of cases were induced by anti-tumour necrosis factor (TNF)-ɑ agents. A median of three drug doses preceded SCLE onset and the lesions lasted a median of 7 weeks after drug cessation. Oral and topical corticosteroids were most frequently used. Of the licensed mAbs, adalimumab, denosumab, rituximab, etanercept and infliximab were calculated to have the highest relative number of yearly users based on global sales data. Comparing the number of mAb-induced SCLE cases with estimated yearly users, the checkpoint inhibitors pembrolizumab and nivolumab showed strikingly high rates of SCLE relative to their global use, but ipilimumab did not.

Conclusion

We present the first systematic review characterising mAb-induced SCLE with respect to triggers, clinical signs, laboratory findings, prognosis and treatment approaches. We identify elevated rates associated with the use of checkpoint inhibitors and anti-TNFɑ agents.

Electronic supplementary material

The online version of this article (10.1007/s40268-020-00320-5) contains supplementary material, which is available to authorized users.

Identification of vascular dementia and Alzheimer's disease hub genes expressed in the frontal lobe and temporal cortex by weighted co-expression network analysis and construction of a protein-protein interaction

Background: It is difficult to distinguish cognitive decline due to AD from that sustained by cerebrovascular disease in view of the great overlap. It is uncertain in the molecular biological pathway behind AD and VaD.Objective: Our study aimed to explore the hub molecules and their associations with each other to identify potential biomarkers and therapeutic targets for the AD and VaD.Methods: We screened the differentially expressed genes of AD and VaD, used weighted gene co-expression network analysis and then constructed a VaD-AD-specific protein-protein interaction network with functional annotation to their related metabolic pathways. Finally, we performed a ROC curve analysis of hub proteins to get an idea about their diagnostic value.Results: In the frontal lobe and temporal cortex, hub genes were identified. With regard to VaD, there were only three hub genes which encoded the neuropeptides, SST, NMU and TAC1. The AUC of these genes were 0.804, 0.768 and 0.779, respectively. One signature was established for these three hub genes with AUC of 0.990. For the identification of AD and VaD, all hub genes were receptors. These genes included SH3GL2, PROK2, TAC3, HTR2A, MET, TF, PTH2R CNR1, CHRM4, PTPN3 and CRH. The AUC of these genes were 0.853, 0.859, 0.796, 0.775, 0.706, 0.677, 0.696, 0.668 and 0.652, respectively. The other signature was built for eleven hub genes with AUC of 0.990.Conclusion: In the frontal lobe and temporal cortex regions, hub genes are used as diagnostic markers, which may provide insight into personalized potential biomarkers and therapeutic targets for patients with VaD and AD.

Differences in substrate use linked to divergent carbon flow during litter decomposition

Discovering widespread microbial processes that create variation in soil carbon (C) cycling within ecosystems may improve soil C modeling. Toward this end, we screened 206 soil communities decomposing plant litter in a common garden microcosm environment and examined features linked to divergent patterns of C flow. C flow was measured as carbon dioxide (CO2) and dissolved organic carbon (DOC) from 44-days of litter decomposition. Two large groups of microbial communities representing ‘high’ and ‘low’ DOC phenotypes from original soil and 44-day microcosm samples were down-selected for fungal and bacterial profiling. Metatranscriptomes were also sequenced from a smaller subset of communities in each group. The two groups exhibited differences in average rate of CO2 production, demonstrating that the divergent patterns of C flow arose from innate functional constraints on C metabolism, not a time-dependent artefact. To infer functional constraints, we identified features – traits at the organism, pathway or gene level – linked to the high and low DOC phenotypes using RNA-Seq approaches and machine learning approaches. Substrate use differed across the high and low DOC phenotypes. Additional features suggested that divergent patterns of C flow may be driven in part by differences in organism interactions that affect DOC abundance directly or indirectly by controlling community structure.

Validity of the cold pressor test and pain sensitivity questionnaire via online self-administration

To determine the feasibility of complex home-based phenotyping, 1,876 research participants from the customer base of 23andMe completed an online version of a Pain Sensitivity Questionnaire (PSQ) as well as a cold pressor test (CPT) which is used in clinical assessments of pain. Overall our online version of the PSQ performed similarly to the original pen-and-paper version. Construct validity of the PSQ total was demonstrated by internal consistency and consistent discrimination between more and less painful items. Criterion validity was demonstrated by correlation with pain sensitivity as measured by the CPT. Within the same cohort we performed a cold pressor test using a layperson description and household equipment. Comparison with published reports from controlled studies revealed similar distributions of cold pain tolerance times (i.e., time elapsed before removing the hand from the water). Of those who elected to participate in the CPT, a large majority of participants did not report issues with the test procedure or noncompliance with the instructions (97%). We confirmed a large sex difference in CPT thresholds in line with published data, such that women removed their hands from the water at a median of 54.2 seconds, with men lasting for a median time of 82.7 seconds (Kruskal-Wallis statistic, p < 0.0001), but other factors like age or current pain treatment were at most weakly associated, and inconsistently between men and women. We introduce a new paradigm for performing pain testing, called testing@home, that, in the case of cold nociception, showed comparable results to studies conducted under controlled conditions and supervision of a health care professional.

Artificial Intelligence Pipeline to Bridge the Gap between Bench Researchers and Clinical Researchers in Precision Medicine

Precision medicine informatics is a field of research that incorporates learning systems that generate new knowledge to improve individualized treatments using integrated data sets and models. Given the ever-increasing volumes of data that are relevant to patient care, artificial intelligence (AI) pipelines need to be a central component of such research to speed discovery. Applying AI methodology to complex multidisciplinary information retrieval can support efforts to discover bridging concepts within collaborating communities. This dovetails with precision medicine research, given the information rich multi-omic data that are used in precision medicine analysis pipelines. In this perspective article we define a prototype AI pipeline to facilitate discovering research connections between bioinformatics and clinical researchers. We propose building knowledge representations that are iteratively improved through AI and human-informed learning feedback loops supported through crowdsourcing. To illustrate this, we will explore the specific use case of nonalcoholic fatty liver disease, a growing health care problem. We will examine AI pipeline construction and utilization in relation to bench-to-bedside bridging concepts with interconnecting knowledge representations applicable to bioinformatics researchers and clinicians.

Deep Phenotyping of Parkinson's Disease

Phenotype is the set of observable traits of an organism or condition. While advances in genetics, imaging, and molecular biology have improved our understanding of the underlying biology of Parkinson's disease (PD), clinical phenotyping of PD still relies primarily on history and physical examination. These subjective, episodic, categorical assessments are valuable for diagnosis and care but have left gaps in our understanding of the PD phenotype. Sensors can provide objective, continuous, real-world data about the PD clinical phenotype, increase our knowledge of its pathology, enhance evaluation of therapies, and ultimately, improve patient care. In this paper, we explore the concept of deep phenotyping-the comprehensive assessment of a condition using multiple clinical, biological, genetic, imaging, and sensor-based tools-for PD. We discuss the rationale for, outline current approaches to, identify benefits and limitations of, and consider future directions for deep clinical phenotyping.

Data-driven method to enhance craniofacial and oral phenotype vocabularies

BACKGROUND

A significant amount of clinical information captured as free-text narratives could be better used for several applications, such as clinical decision support, ontology development, evidence-based practice, and research. The Human Phenotype Ontology (HPO) is specifically used for semantic comparisons for diagnostic purposes. All these functions require quality coverage of the domain of interest. The authors used natural language processing to capture craniofacial and oral phenotype signatures from electronic health records and then used these signatures for evaluation of existing oral phenotype ontology coverage.

METHODS

The authors applied a text-processing pipeline based on the clinical Text Analysis and Knowledge Extraction System to annotate the clinical notes with Unified Medical Language System codes. The authors extracted the disease or disorder phenotype terms, which were then compared with HPO terms and their synonyms.

RESULTS

The authors retrieved 2,153 deidentified clinical notes from 558 patients. Finally, 2,416 unique diseases or disorders phenotype terms were extracted, which included 210 craniofacial or oral phenotype terms. Twenty-six of these phenotypes were not found in the HPO.

CONCLUSIONS

The authors demonstrated that natural language processing tools could extract relevant phenotype terms from clinical narratives, which could help identify gaps in existing ontologies and enhance craniofacial and dental phenotyping vocabularies.

PRACTICAL IMPLICATIONS

The expansion of terms in the dental, oral, and craniofacial domains in the HPO is particularly important as the dental community moves toward electronic health records.

Enabling Global Clinical Collaborations on Identifiable Patient Data: The Minerva Initiative

The clinical utility of computational phenotyping for both genetic and rare diseases is increasingly appreciated; however, its true potential is yet to be fully realized. Alongside the growing clinical and research availability of sequencing technologies, precise deep and scalable phenotyping is required to serve unmet need in genetic and rare diseases. To improve the lives of individuals affected with rare diseases through deep phenotyping, global big data interrogation is necessary to aid our understanding of disease biology, assist diagnosis, and develop targeted treatment strategies. This includes the application of cutting-edge machine learning methods to image data. As with most digital tools employed in health care, there are ethical and data governance challenges associated with using identifiable personal image data. There are also risks with failing to deliver on the patient benefits of these new technologies, the biggest of which is posed by data siloing. The Minerva Initiative has been designed to enable the public good of deep phenotyping while mitigating these ethical risks. Its open structure, enabling collaboration and data sharing between individuals, clinicians, researchers and private enterprise, is key for delivering precision public health.

Exploring the dark genome: implications for precision medicine

The increase in the number of both patients and healthcare practitioners who grew up using the Internet and computers (so-called "digital natives") is likely to impact the practice of precision medicine, and requires novel platforms for data integration and mining, as well as contextualized information retrieval. The "Illuminating the Druggable Genome Knowledge Management Center" (IDG KMC) quantifies data availability from a wide range of chemical, biological, and clinical resources, and has developed platforms that can be used to navigate understudied proteins (the "dark genome"), and their potential contribution to specific pathologies. Using the "Target Importance and Novelty Explorer" (TIN-X) highlights the role of LRRC10 (a dark gene) in dilated cardiomyopathy. Combining mouse and human phenotype data leads to increased strength of evidence, which is discussed for four additional dark genes: SLX4IP and its role in glucose metabolism, the role of HSF2BP in coronary artery disease, the involvement of ELFN1 in attention-deficit hyperactivity disorder and the role of VPS13D in mouse neural tube development and its confirmed role in childhood onset movement disorders. The workflow and tools described here are aimed at guiding further experimental research, particularly within the context of precision medicine.

Feasibility study of a point-of-care positron emission tomography system with interactive imaging capability

PURPOSE

We investigated the feasibility of a novel positron emission tomography (PET) system that provides near real-time feedback to an operator who can interactively scan a patient to optimize image quality. The system should be compact and mobile to support point-of-care (POC) molecular imaging applications. In this study, we present the key technologies required and discuss the potential benefits of such new capability.

METHODS

The core of this novel PET technology includes trackable PET detectors and a fully three-dimensional, fast image reconstruction engine implemented on multiple graphics processing units (GPUs) to support dynamically changing geometry by calculating the system matrix on-the-fly using a tube-of-response approach. With near real-time image reconstruction capability, a POC-PET system may comprise a maneuverable front PET detector and a second detector panel which can be stationary or moved synchronously with the front detector such that both panels face the region-of-interest (ROI) with the detector trajectory contoured around a patient's body. We built a proof-of-concept prototype using two planar detectors each consisting of a photomultiplier tube (PMT) optically coupled to an array of 48 × 48 lutetium-yttrium oxyorthosilicate (LYSO) crystals (1.0 × 1.0 × 10.0 mm³ each). Only 38 × 38 crystals in each arrays can be clearly re-solved and used for coincidence detection. One detector was mounted to a robotic arm which can position it at arbitrary locations, and the other detector was mounted on a rotational stage. A cylindrical phantom (102 mm in diameter, 150 mm long) with nine spherical lesions (8:1 tumor-to-background activity concentration ratio) was imaged from 27 sampling angles. List-mode events were reconstructed to form images without or with time-of-flight (TOF) information. We conducted two Monte Carlo simulations using two POC-PET systems. The first one uses the same phantom and detector setup as our experiment, with the detector coincidence re-solving time (CRT) ranging from 100 to 700 ps full-width-at-half-maximum (FWHM). The second study simulates a body-size phantom (316 × 228 × 160 mm³ ) imaged by a larger POC-PET system that has 4 × 6 modules (32 × 32 LYSO crystals/module, four in axial and six in transaxial directions) in the front panel and 3 × 8 modules (16 × 16 LYSO crystals/module, three in axial and eight in transaxial directions) in the back panel. We also evaluated an interactive scanning strategy by progressively increasing the number of data sets used for image reconstruction. The updated images were analyzed based on the number of data sets and the detector CRT.

RESULTS

The proof-of-concept prototype re-solves most of the spherical lesions despite a limited number of coincidence events and incomplete sampling. TOF information reduces artifacts in the reconstructed images. Systems with better timing resolution exhibit improved image quality and reduced artifacts. We observed a reconstruction speed of 0.96 × 10⁶ events/s/iteration for 600 × 600 × 224 voxel rectilinear space using four GPUs. A POC-PET system with significantly higher sensitivity can interactively image a body-size object from four angles in less than 7 min.

CONCLUSIONS

We have developed GPU-based fast image reconstruction capability to support a PET system with arbitrary and dynamically changing geometry. Using TOF PET detectors, we demonstrated the feasibility of a PET system that can provide timely visual feedback to an operator who can scan a patient interactively to support POC imaging applications.

Merging heterogeneous clinical data to enable knowledge discovery

The vision of precision medicine relies on the integration of large-scale clinical, molecular and environmental datasets. Data integration may be thought of along two axes: data fusion across institutions, and data fusion across modalities. Cross-institutional data sharing that maintains semantic integrity hinges on the adoption of data standards and a push toward ontology-driven integration. The goal should be the creation of query-able data repositories spanning primary and tertiary care providers, disease registries, research organizations etc. to produce rich longitudinal datasets. Cross-modality sharing involves the integration of multiple data streams, from structured EHR data (diagnosis codes, laboratory tests) to genomics, imaging, monitors and patient-generated data including wearable devices. This integration presents unique technical, semantic, and ethical challenges; however recent work suggests that multi-modal clinical data can significantly improve the performance of phenotyping and prediction algorithms, powering knowledge discovery at the patient- and population-level.

Phelan-McDermid syndrome data network: Integrating patient reported outcomes with clinical notes and curated genetic reports

The heterogeneity of patient phenotype data are an impediment to the research into the origins and progression of neuropsychiatric disorders. This difficulty is compounded in the case of rare disorders such as Phelan-McDermid Syndrome (PMS) by the paucity of patient clinical data. PMS is a rare syndromic genetic cause of autism and intellectual deficiency. In this paper, we describe the Phelan-McDermid Syndrome Data Network (PMS_DN), a platform that facilitates research into phenotype-genotype correlation and progression of PMS by: a) integrating knowledge of patient phenotypes extracted from Patient Reported Outcomes (PRO) data and clinical notes-two heterogeneous, underutilized sources of knowledge about patient phenotypes-with curated genetic information from the same patient cohort and b) making this integrated knowledge, along with a suite of statistical tools, available free of charge to authorized investigators on a Web portal https://pmsdn.hms.harvard.edu. PMS_DN is a Patient Centric Outcomes Research Initiative (PCORI) where patients and their families are involved in all aspects of the management of patient data in driving research into PMS. To foster collaborative research, PMS_DN also makes patient aggregates from this knowledge available to authorized investigators using distributed research networks such as the PCORnet PopMedNet. PMS_DN is hosted on a scalable cloud based environment and complies with all patient data privacy regulations. As of October 31, 2016, PMS_DN integrates high-quality knowledge extracted from the clinical notes of 112 patients and curated genetic reports of 176 patients with preprocessed PRO data from 415 patients.

The anatomy of phenotype ontologies: principles, properties and applications

The past decade has seen an explosion in the collection of genotype data in domains as diverse as medicine, ecology, livestock and plant breeding. Along with this comes the challenge of dealing with the related phenotype data, which is not only large but also highly multidimensional. Computational analysis of phenotypes has therefore become critical for our ability to understand the biological meaning of genomic data in the biological sciences. At the heart of computational phenotype analysis are the phenotype ontologies. A large number of these ontologies have been developed across many domains, and we are now at a point where the knowledge captured in the structure of these ontologies can be used for the integration and analysis of large interrelated data sets. The Phenotype And Trait Ontology framework provides a method for formal definitions of phenotypes and associated data sets and has proved to be key to our ability to develop methods for the integration and analysis of phenotype data. Here, we describe the development and products of the ontological approach to phenotype capture, the formal content of phenotype ontologies and how their content can be used computationally.

Emerging Role of Precision Medicine in Cardiovascular Disease

Precision medicine is an integrative approach to cardiovascular disease prevention and treatment that considers an individual's genetics, lifestyle, and exposures as determinants of their cardiovascular health and disease phenotypes. This focus overcomes the limitations of reductionism in medicine, which presumes that all patients with the same signs of disease share a common pathophenotype and, therefore, should be treated similarly. Precision medicine incorporates standard clinical and health record data with advanced panomics (ie, transcriptomics, epigenomics, proteomics, metabolomics, and microbiomics) for deep phenotyping. These phenotypic data can then be analyzed within the framework of molecular interaction (interactome) networks to uncover previously unrecognized disease phenotypes and relationships between diseases, and to select pharmacotherapeutics or identify potential protein-drug or drug-drug interactions. In this review, we discuss the current spectrum of cardiovascular health and disease, population averages and the response of extreme phenotypes to interventions, and population-based versus high-risk treatment strategies as a pretext to understanding a precision medicine approach to cardiovascular disease prevention and therapeutic interventions. We also consider the search for resilience and Mendelian disease genes and argue against the theory of a single causal gene/gene product as a mediator of the cardiovascular disease phenotype, as well as an Erlichian magic bullet to solve cardiovascular disease. Finally, we detail the importance of deep phenotyping and interactome networks and the use of this information for rational polypharmacy. These topics highlight the urgent need for precise phenotyping to advance precision medicine as a strategy to improve cardiovascular health and prevent disease.

PhenoLines: Phenotype Comparison Visualizations for Disease Subtyping via Topic Models

PhenoLines is a visual analysis tool for the interpretation of disease subtypes, derived from the application of topic models to clinical data. Topic models enable one to mine cross-sectional patient comorbidity data (e.g., electronic health records) and construct disease subtypes-each with its own temporally evolving prevalence and co-occurrence of phenotypes-without requiring aligned longitudinal phenotype data for all patients. However, the dimensionality of topic models makes interpretation challenging, and de facto analyses provide little intuition regarding phenotype relevance or phenotype interrelationships. PhenoLines enables one to compare phenotype prevalence within and across disease subtype topics, thus supporting subtype characterization, a task that involves identifying a proposed subtype's dominant phenotypes, ages of effect, and clinical validity. We contribute a data transformation workflow that employs the Human Phenotype Ontology to hierarchically organize phenotypes and aggregate the evolving probabilities produced by topic models. We introduce a novel measure of phenotype relevance that can be used to simplify the resulting topology. The design of PhenoLines was motivated by formative interviews with machine learning and clinical experts. We describe the collaborative design process, distill high-level tasks, and report on initial evaluations with machine learning experts and a medical domain expert. These results suggest that PhenoLines demonstrates promising approaches to support the characterization and optimization of topic models.

An ontology-aware integration of clinical models, terminologies and guidelines: an exploratory study of the Scale for the Assessment and Rating of Ataxia (SARA)

Background

Electronic rating scales represent an important resource for standardized data collection. However, the ability to exploit reasoning on rating scale data is still limited. The objective of this work is to facilitate the integration of the semantics required to automatically interpret collections of standardized clinical data. We developed an electronic prototype for the Scale of the Assessment and Rating of Ataxia (SARA), broadly used in neurology. In order to address the modeling challenges of the SARA, we propose to combine the best performances from OpenEHR clinical archetypes, guidelines and ontologies.

Methods

A scaled-down version of the Human Phenotype Ontology (HPO) was built, extracting the terms that describe the SARA tests from free-text sources. This version of the HPO was then used as backbone to normalize the content of the SARA through clinical archetypes. The knowledge required to exploit reasoning on the SARA data was modeled as separate information-processing units interconnected via the defined archetypes. Each unit used the most appropriate technology to formally represent the required knowledge.

Results

Based on this approach, we implemented a prototype named SARA Management System, to be used for both the assessment of cerebellar syndrome and the production of a clinical synopsis. For validation purposes, we used recorded SARA data from 28 anonymous subjects affected by Spinocerebellar Ataxia Type 36 (SCA36). When comparing the performance of our prototype with that of two independent experts, weighted kappa scores ranged from 0.62 to 0.86.

Conclusions

The combination of archetypes, phenotype ontologies and electronic information-processing rules can be used to automate the extraction of relevant clinical knowledge from plain scores of rating scales. Our results reveal a substantial degree of agreement between the results achieved by an ontology-aware system and the human experts.

Electronic supplementary material

The online version of this article (10.1186/s12911-017-0568-4) contains supplementary material, which is available to authorized users.

Use of Biomedical Ontologies for Integration of Biological Knowledge for Learning and Prediction of Adverse Drug Reactions

Drug-induced toxicity is a major public health concern that leads to patient morbidity and mortality. To address this problem, the Food and Drug Administration is working on the PredicTox initiative, a pilot research program on tyrosine kinase inhibitors, to build mechanistic and predictive models for drug-induced toxicity. This program involves integrating data acquired during preclinical studies and clinical trials within pharmaceutical company development programs that they have agreed to put in the public domain and in publicly available biological, pharmacological, and chemical databases. The integration process is accommodated by biomedical ontologies, a set of standardized vocabularies that define terms and logical relationships between them in each vocabulary. We describe a few programs that have used ontologies to address biomedical questions. The PredicTox effort is leveraging the experience gathered from these early initiatives to develop an infrastructure that allows evaluation of the hypothesis that having a mechanistic understanding underlying adverse drug reactions will improve the capacity to understand drug-induced clinical adverse drug reactions.

PhenoStacks: Cross-Sectional Cohort Phenotype Comparison Visualizations

Cross-sectional phenotype studies are used by genetics researchers to better understand how phenotypes vary across patients with genetic diseases, both within and between cohorts. Analyses within cohorts identify patterns between phenotypes and patients (e.g., co-occurrence) and isolate special cases (e.g., potential outliers). Comparing the variation of phenotypes between two cohorts can help distinguish how different factors affect disease manifestation (e.g., causal genes, age of onset, etc.). PhenoStacks is a novel visual analytics tool that supports the exploration of phenotype variation within and between cross-sectional patient cohorts. By leveraging the semantic hierarchy of the Human Phenotype Ontology, phenotypes are presented in context, can be grouped and clustered, and are summarized via overviews to support the exploration of phenotype distributions. The design of PhenoStacks was motivated by formative interviews with genetics researchers: we distil high-level tasks, present an algorithm for simplifying ontology topologies for visualization, and report the results of a deployment evaluation with four expert genetics researchers. The results suggest that PhenoStacks can help identify phenotype patterns, investigate data quality issues, and inform data collection design.

The Human Phenotype Ontology in 2017

Deep phenotyping has been defined as the precise and comprehensive analysis of phenotypic abnormalities in which the individual components of the phenotype are observed and described. The three components of the Human Phenotype Ontology (HPO; www.human-phenotype-ontology.org) project are the phenotype vocabulary, disease-phenotype annotations and the algorithms that operate on these. These components are being used for computational deep phenotyping and precision medicine as well as integration of clinical data into translational research. The HPO is being increasingly adopted as a standard for phenotypic abnormalities by diverse groups such as international rare disease organizations, registries, clinical labs, biomedical resources, and clinical software tools and will thereby contribute toward nascent efforts at global data exchange for identifying disease etiologies. This update article reviews the progress of the HPO project since the debut Nucleic Acids Research database article in 2014, including specific areas of expansion such as common (complex) disease, new algorithms for phenotype driven genomic discovery and diagnostics, integration of cross-species mapping efforts with the Mammalian Phenotype Ontology, an improved quality control pipeline, and the addition of patient-friendly terminology.