Deep Phenotypic Analysis for Transposition of the Great Arteries and Prognosis Implication

Background Transposition of the great arteries (TGA) consists of about 3% of all congenital heart diseases and 20% of cyanotic congenital heart diseases. It is always accompanied by a series of other cardiac malformations that affect the surgical intervention strategy as well as prognosis. In this study, we comprehensively analyzed the phenotypes of the patients who had TGA with concordant atrioventricular and discordant ventriculoarterial connections and explored their association with prognosis. Methods and Results We retrospectively reviewed 666 patients with a diagnosis of TGA with concordant atrioventricular and discordant ventriculoarterial connections in Fuwai Hospital from 1997 to 2019. Under the guidance of the Human Phenotype Ontology database, patients were classified into 3 clusters. The Kaplan‐Meier method was used to analyze the prognosis, and the Cox proportional regression model was used to investigate the risk factors. In this 666‐patient TGA cohort, the overall 5‐year survival rate was 94.70% (92.95%–96.49%). Three clusters with distinct phenotypes were obtained by the Human Phenotype Ontology database. Kaplan‐Meier analysis revealed a significant difference in freedom from reintervention among 3 clusters ( P <0.001). To eliminate the effect of surgeries, we analyzed patients who only received an arterial switch operation and still found a significant difference in reintervention ( P =0.019). Conclusions We delineated a big cardiovascular phenotypic profile of an unprecedentedly large TGA cohort and successfully risk stratified them to reveal prognostic significance. Also, we reported the outcomes of a large TGA population in China.

Social Media Validation for Pre-Screened Ontologies As a Driver for Biomolecular Interaction Significance in COVID-19 and Hematological Malignancies

With the rise of social media use during the COVID-19 pandemic, impressions from online content can affect behavioral changes resulting in exacerbating disparities in care. Thus, there exists a need to utilize social media platforms, like Twitter, to help augment preparedness, especially at the intersection between oncology and COVID-19, where tweets could help hint at potential biomolecular interactions. To address this, a study was developed to assess relationship and ontologies on the interaction between hematological malignancies and COVID-19 on Twitter. Ontologies are groupings of terms and related identifiers, such as genes, for general search terms, such as "Blood Cancer", were found utilizing the Human Phenotype Ontology. These were combined with the term "COVID-19" and used as search terms for Twitter's Standard Search API. The resulting tweets were cross-checked to assess if they included any of the other terms or genes related to the starting ontologies to then determine how many terms or genes each tweet was associated with. Once the most associated tweets to the ontologies were found, the genes related to those ontologies were utilized to find biological structures within the AlphaFold EMBL database, before being used in binding using HEX Docking software's shape based binding tool in 3D. Finally, Root Mean Square (RMS) Deviations were performed between the top 2000 conformations for each bound structure to determine if the binding was statistically significant. Results showed strong clustering of top tweets around keyword combinations. In the case of the starting entry, "Blood COVID-19", the ontologies that were found were linked to 45 terms that each had 100 or more tweets linked to them (Figure 1a). One such term of significance was Acute Myeloid Leukemia, which was linked to the gene BRCA1. The biological significance of the molecular interaction between BRCA1 and SARS CoV-2 was determined using the predicted protein structure from the AlphaFold-EMBL database for BRCA1 and the RCSB Protein Bank structure for the SARS CoV-2 spike (PDB# 6VSB), which can be found in Figure 1b. This interaction was found to be significant based on the average RMS Deviation of 82.97 Angstroms that ranged across the top 2000 conformation. Each model had an average RMS of 85.05 Angstroms between BRCA1 and the COVID-19 spike, with binding occurring on the spike's carbohydrate recognition domain within its S1 segment that is typically used for cell entry. Thus, human phenotype ontology was effective in classifying tweets to specific biomolecular interactions. Therefore, this approach could be utilized to proactively influence treatment designs for blood cancer patients infected with COVID-19, as well as in other areas where medical illnesses are already well defined by ontologies or other literature data. Forward looking, future studies will help to ensure that terms that are not well characterized by ontologies can still be utilized in this type of analysis by employing de novo ontology production methods. Figure 1 Figure 1. Disclosures No relevant conflicts of interest to declare.

PhenoTagger: a hybrid method for phenotype concept recognition using human phenotype ontology

Motivation Automatic phenotype concept recognition from unstructured text remains a challenging task in biomedical text mining research. Previous works that address the task typically use dictionary-based matching methods, which can achieve high precision but suffer from lower recall. Recently, machine learning-based methods have been proposed to identify biomedical concepts, which can recognize more unseen concept synonyms by automatic feature learning. However, most methods require large corpora of manually annotated data for model training, which is difficult to obtain due to the high cost of human annotation. Results In this article, we propose PhenoTagger, a hybrid method that combines both dictionary and machine learning-based methods to recognize Human Phenotype Ontology (HPO) concepts in unstructured biomedical text. We first use all concepts and synonyms in HPO to construct a dictionary, which is then used to automatically build a distantly supervised training dataset for machine learning. Next, a cutting-edge deep learning model is trained to classify each candidate phrase (n-gram from input sentence) into a corresponding concept label. Finally, the dictionary and machine learning-based prediction results are combined for improved performance. Our method is validated with two HPO corpora, and the results show that PhenoTagger compares favorably to previous methods. In addition, to demonstrate the generalizability of our method, we retrained PhenoTagger using the disease ontology MEDIC for disease concept recognition to investigate the effect of training on different ontologies. Experimental results on the NCBI disease corpus show that PhenoTagger without requiring manually annotated training data achieves competitive performance as compared with state-of-the-art supervised methods. Availabilityand implementation The source code, API information and data for PhenoTagger are freely available at https://github.com/ncbi-nlp/PhenoTagger. Supplementary information Supplementary data are available at Bioinformatics online.

The Human Phenotype Ontology in 2021

The Human Phenotype Ontology (HPO, https://hpo.jax.org) was launched in 2008 to provide a comprehensive logical standard to describe and computationally analyze phenotypic abnormalities found in human disease. The HPO is now a worldwide standard for phenotype exchange. The HPO has grown steadily since its inception due to considerable contributions from clinical experts and researchers from a diverse range of disciplines. Here, we present recent major extensions of the HPO for neurology, nephrology, immunology, pulmonology, newborn screening, and other areas. For example, the seizure subontology now reflects the International League Against Epilepsy (ILAE) guidelines and these enhancements have already shown clinical validity. We present new efforts to harmonize computational definitions of phenotypic abnormalities across the HPO and multiple phenotype ontologies used for animal models of disease. These efforts will benefit software such as Exomiser by improving the accuracy and scope of cross-species phenotype matching. The computational modeling strategy used by the HPO to define disease entities and phenotypic features and distinguish between them is explained in detail.We also report on recent efforts to translate the HPO into indigenous languages. Finally, we summarize recent advances in the use of HPO in electronic health record systems.