Biological applications of knowledge graph embedding models.

Complex biological systems are traditionally modelled as graphs of interconnected biological entities. These graphs, i.e. biological knowledge graphs, are then processed using graph exploratory approaches to perform different types of analytical and predictive tasks. Despite the high predictive accuracy of these approaches, they have limited scalability due to their dependency on time-consuming path exploratory procedures. In recent years, owing to the rapid advances of computational technologies, new approaches for modelling graphs and mining them with high accuracy and scalability have emerged. These approaches, i.e. knowledge graph embedding (KGE) models, operate by learning low-rank vector representations of graph nodes and edges that preserve the graph's inherent structure. These approaches were used to analyse knowledge graphs from different domains where they showed superior performance and accuracy compared to previous graph exploratory approaches. In this work, we study this class of models in the context of biological knowledge graphs and their different applications. We then show how KGE models can be a natural fit for representing complex biological knowledge modelled as graphs. We also discuss their predictive and analytical capabilities in different biology applications. In this regard, we present two example case studies that demonstrate the capabilities of KGE models: prediction of drug-target interactions and polypharmacy side effects. Finally, we analyse different practical considerations for KGEs, and we discuss possible opportunities and challenges related to adopting them for modelling biological systems.

Discovering protein drug targets using knowledge graph embeddings.

MOTIVATION: Computational approaches for predicting drug-target interactions (DTIs) can provide valuable insights into the drug mechanism of action. DTI predictions can help to quickly identify new promising (on-target) or unintended (off-target) effects of drugs. However, existing models face several challenges. Many can only process a limited number of drugs and/or have poor proteome coverage. The current approaches also often suffer from high false positive prediction rates. RESULTS: We propose a novel computational approach for predicting drug target proteins. The approach is based on formulating the problem as a link prediction in knowledge graphs (robust, machine-readable representations of networked knowledge). We use biomedical knowledge bases to create a knowledge graph of entities connected to both drugs and their potential targets. We propose a specific knowledge graph embedding model, TriModel, to learn vector representations (i.e. embeddings) for all drugs and targets in the created knowledge graph. These representations are consequently used to infer candidate drug target interactions based on their scores computed by the trained TriModel model. We have experimentally evaluated our method using computer simulations and compared it to five existing models. This has shown that our approach outperforms all previous ones in terms of both area under ROC and precision-recall curves in standard benchmark tests. AVAILABILITY AND IMPLEMENTATION: The data, predictions and models are available at: drugtargets.insight-centre.org. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Accurate prediction of kinase-substrate networks using knowledge graphs.

Phosphorylation of specific substrates by protein kinases is a key control mechanism for vital cell-fate decisions and other cellular processes. However, discovering specific kinase-substrate relationships is time-consuming and often rather serendipitous. Computational predictions alleviate these challenges, but the current approaches suffer from limitations like restricted kinome coverage and inaccuracy. They also typically utilise only local features without reflecting broader interaction context. To address these limitations, we have developed an alternative predictive model. It uses statistical relational learning on top of phosphorylation networks interpreted as knowledge graphs, a simple yet robust model for representing networked knowledge. Compared to a representative selection of six existing systems, our model has the highest kinome coverage and produces biologically valid high-confidence predictions not possible with the other tools. Specifically, we have experimentally validated predictions of previously unknown phosphorylations by the LATS1, AKT1, PKA and MST2 kinases in human. Thus, our tool is useful for focusing phosphoproteomic experiments, and facilitates the discovery of new phosphorylation reactions. Our model can be accessed publicly via an easy-to-use web interface (LinkPhinder).

ACURATE neo2 Transcatheter aortic valve implantation without balloon aortic valvuloplasty - direct ACURATE neo2.

BACKGROUND: ACURATE neo2 (Neo2) implantation is performed after systematic Balloon Aortic Valvuloplasty (BAV) in most patients. No reports exist about the feasibility and safety of direct Neo2 transcatheter aortic valve implantation (TAVI) in comparison to the standard practice. AIM: We aimed to identify the patients' baseline anatomical characteristics, procedural, and early post-procedural outcomes in patients treated using Neo2 with and without BAV. METHODOLOGY: This is a retrospective multicentre analysis of 499 patients with severe aortic stenosis who underwent TAVI using Neo2. The comparison was done according to the performance or omission of BAV. Echocardiography and computed tomography were analysed by an independent Core Lab. Propensity score matching (PSM) was performed based on the annular diameter and AV calcium volume, which identified 84 matched pairs. RESULTS: Among the cohort included, 391 (78%) patients received BAV (BAV-yes) and 108 (22%) were not attempted (BAV-no or Direct TAVI). Patients in BAV-no cohort had smaller annular diameter (22.6 vs 23.4 mm; p < 0.001) and lower calcium volume (163 vs 581 mm3; p < 0.001) compared to BAV-yes cohort. In the matched cohort, VARC-3 device technical success was similar (95%) and all other outcome measures were statistically comparable between cohorts. CONCLUSION: Direct TAVI using ACURATEneo2 without pre-TAVI balloon aortic valvuloplasty in patients with mild or less valve calcifications might be feasible and associated with comparable early outcomes compared to patients with similar anatomical features undergoing systematic balloon valvuloplasty.

Predictors of Conduction Disturbances Requiring New Permanent Pacemaker Implantation following Transcatheter Aortic Valve Implantation Using the Evolut Series.

(1) Background: Conduction disturbance requiring a new permanent pacemaker (PPM) after transcatheter aortic valve implantation (TAVI) has traditionally been a common complication. New implantation techniques with self-expanding platforms have reportedly reduced the incidence of PPM. We sought to investigate the predictors of PPM at 30 days after TAVI using Evolut R/PRO/PRO+; (2) Methods: Consecutive patients who underwent TAVI with the Evolut platform between October 2019 and August 2022 at University Hospital Galway, Ireland, were included. Patients who had a prior PPM (n = 10), valve-in-valve procedures (n = 8) or received >1 valve during the index procedure (n = 3) were excluded. Baseline clinical, electrocardiographic (ECG), echocardiographic and multislice computed tomography (MSCT) parameters were analyzed. Pre-TAVI MSCT analysis included membranous septum (MS) length, a semi-quantitative calcification analysis of the aortic valve leaflets, left ventricular outflow tract, and mitral annulus. Furthermore, the implantation depth (ID) was measured from the final aortography. Multivariate binary logistic analysis and receiver operating characteristic (ROC) curve analysis were used to identify independent predictors and the optimal MS and ID cutoff values to predict new PPM requirements, respectively; (3) Results: A total of 129 TAVI patients were included (age = 81.3 ± 5.3 years; 36% female; median EuroSCORE II 3.2 [2.0, 5.4]). Fifteen patients (11.6%) required PPM after 30 days. The patients requiring new PPM at 30 days were more likely to have a lower European System for Cardiac Operative Risk Evaluation II, increased prevalence of right bundle branch block (RBBB) at baseline ECG, have a higher mitral annular calcification severity and have a shorter MS on preprocedural MSCT analysis, and have a ID, as shown on the final aortogram. From the multivariate analysis, pre-TAVI RBBB, MS length, and ID were shown to be predictors of new PPM. An MS length of <2.85 mm (AUC = 0.85, 95%CI: (0.77, 0.93)) and ID of >3.99 mm (area under the curve (AUC) = 0.79, (95% confidence interval (CI): (0.68, 0.90)) were found to be the optimal cut-offs for predicting new PPM requirements; (4) Conclusions: Membranous septum length and implantation depth were found to be independent predictors of new PPM post-TAVI with the Evolut platform. Patient-specific implantation depth could be used to mitigate the requirement for new PPM.

Machine Learning-Assisted Recurrence Prediction for Patients With Early-Stage Non-Small-Cell Lung Cancer.

PURPOSE: Stratifying patients with cancer according to risk of relapse can personalize their care. In this work, we provide an answer to the following research question: How to use machine learning to estimate probability of relapse in patients with early-stage non-small-cell lung cancer (NSCLC)? MATERIALS AND METHODS: For predicting relapse in 1,387 patients with early-stage (I-II) NSCLC from the Spanish Lung Cancer Group data (average age 65.7 years, female 24.8%, male 75.2%), we train tabular and graph machine learning models. We generate automatic explanations for the predictions of such models. For models trained on tabular data, we adopt SHapley Additive exPlanations local explanations to gauge how each patient feature contributes to the predicted outcome. We explain graph machine learning predictions with an example-based method that highlights influential past patients. RESULTS: Machine learning models trained on tabular data exhibit a 76% accuracy for the random forest model at predicting relapse evaluated with a 10-fold cross-validation (the model was trained 10 times with different independent sets of patients in test, train, and validation sets, and the reported metrics are averaged over these 10 test sets). Graph machine learning reaches 68% accuracy over a held-out test set of 200 patients, calibrated on a held-out set of 100 patients. CONCLUSION: Our results show that machine learning models trained on tabular and graph data can enable objective, personalized, and reproducible prediction of relapse and, therefore, disease outcome in patients with early-stage NSCLC. With further prospective and multisite validation, and additional radiological and molecular data, this prognostic model could potentially serve as a predictive decision support tool for deciding the use of adjuvant treatments in early-stage lung cancer.

Core Lab Adjudication of the ACURATE neo2 Hemodynamic Performance Using Computed-Tomography-Corrected Left Ventricular Outflow Tract Area.

(1) Background: Hemodynamic assessment of prosthetic heart valves using conventional 2D transthoracic Echocardiography-Doppler (2D-TTE) has limitations. Of those, left ventricular outflow tract (LVOT) area measurement is one of the major limitations of the continuity equation, which assumes a circular LVOT. (2) Methods: This study comprised 258 patients with severe aortic stenosis (AS), who were treated with the ACURATE neo2. The LVOT area and its dependent Doppler-derived parameters, including effective orifice area (EOA) and stroke volume (SV), in addition to their indexed values, were calculated from post-TAVI 2D-TTE. In addition, the 3D-LVOT area from pre-procedural MDCT scans was obtained and used to calculate corrected Doppler-derived parameters. The incidence rates of prosthesis patient mismatch (PPM) were compared between the 2D-TTE and MDCT-based methods (3) Results: The main results show that the 2D-TTE measured LVOT is significantly smaller than 3D-MDCT (350.4 ± 62.04 mm2 vs. 405.22 ± 81.32 mm2) (95% Credible interval (CrI) of differences: -55.15, -36.09), which resulted in smaller EOA (2.25 ± 0.59 vs. 2.58 ± 0.63 cm2) (Beta = -0.642 (95%CrI of differences: -0.85, -0.43), and lower SV (73.88 ± 21.41 vs. 84.47 ± 22.66 mL), (Beta = -7.29 (95% CrI: -14.45, -0.14)), respectively. PPM incidence appears more frequent with 2D-TTE- than 3D-MDCT-corrected measurements (based on the EOAi) 8.52% vs. 2.32%, respectively. In addition, significant differences regarding the EOA among the three valve sizes (S, M and L) were seen only with the MDCT, but not on 2D-TTE. (4) Conclusions: The corrected continuity equation by combining the 3D-LVOT area from MDCT with the TTE Doppler parameters might provide a more accurate assessment of hemodynamic parameters and PPM diagnosis in patients treated with TAVI. The ACURATE neo2 THV has a large EOA and low incidence of PPM using the 3D-corrected LVOT area than on 2D-TTE. These findings need further confirmation on long-term follow-up and in other studies.

On Predicting Recurrence in Early Stage Non-small Cell Lung Cancer.

Early detection and mitigation of disease recurrence in non-small cell lung cancer (NSCLC) patients is a nontrivial problem that is typically addressed either by rather generic follow-up screening guidelines, self-reporting, simple nomograms, or by models that predict relapse risk in individual patients using statistical analysis of retrospective data. We posit that machine learning models trained on patient data can provide an alternative approach that allows for more efficient development of many complementary models at once, superior accuracy, less dependency on the data collection protocols and increased support for explainability of the predictions. In this preliminary study, we describe an experimental suite of various machine learning models applied on a patient cohort of 2442 early stage NSCLC patients. We discuss the promising results achieved, as well as the lessons we learned while developing this baseline for further, more advanced studies in this area.

Predicting The Effects of Chemical-Protein Interactions On Proteins Using Tensor Factorisation.

Understanding the different effects of chemical substances on human proteins is fundamental for designing new drugs. It is also important for elucidating the different mechanisms of action of drugs that can cause side-effects. In this context, computational methods for predicting chemical-protein interactions can provide valuable insights on the relation between therapeutic chemical substances and proteins. Their predictions therefore can help in multiple tasks such as drug repurposing, identifying new drug side-effects, etc. Despite their useful predictions, these methods are unable to predict the different implications - such as change in protein expression, abundance, etc, - of chemical - protein interactions. Therefore, In this work, we study the modelling of chemical-protein interactions' effects on proteins activity using computational approaches. We hereby propose using 3D tensors to model chemicals, their target proteins and the effects associated to their interactions. We then use multi-part embedding tensor factorisation to predict the different effects of chemicals on human proteins. We assess the predictive accuracy of our proposed method using a benchmark dataset that we built. We then show by computational experimental evaluation that our approach outperforms other tensor factorisation methods in the task of predicting effects of chemicals on human proteins.

Predicting Polypharmacy Side-effects Using Knowledge Graph Embeddings.

Polypharmacy is the use of drug combinations and is commonly used for treating complex and terminal diseases. Despite its effectiveness in many cases, it poses high risks of adverse side effects. Polypharmacy side-effects occur due to unwanted interactions of combined drugs, and they can cause severe complications to patients which results in increasing the risks of morbidity and leading to new mortalities. The use of drug polypharmacy is currently in its early stages; thus, the knowledge of their probable side-effects is limited. This encouraged multiple works to investigate machine learning techniques to efficiently and reliably predict adverse effects of drug combinations. In this context, the Decagon model is known to provide state-of-the-art results. It models polypharmacy side-effect data as a knowledge graph and formulates finding possible adverse effects as a link prediction task over the knowledge graph. The link prediction is solved using an embedding model based on graph convolutions. Despite its effectiveness, the Decagon approach still suffers from a high rate of false positives. In this work, we propose a new knowledge graph embedding technique that uses multi-part embedding vectors to predict polypharmacy side-effects. Like in the Decagon model, we model polypharmacy side effects as a knowledge graph. However, we perform the link prediction task using an approach based on tensor decomposition. Our experimental evaluation shows that our approach outperforms the Decagon model with 12% and 16% margins in terms of the area under the ROC and precision recall curves, respectively.