Development and external validation of a dynamic risk score for early prediction of cardiogenic shock in cardiac intensive care units using machine learning.

BACKGROUND: Myocardial infarction and heart failure are major cardiovascular diseases that affect millions of people in the US with the morbidity and mortality being highest among patients who develop cardiogenic shock. Early recognition of cardiogenic shock allows prompt implementation of treatment measures. Our objective is to develop a new dynamic risk score, called CShock, to improve early detection of cardiogenic shock in cardiac intensive care unit (ICU). METHODS: We developed and externally validated a deep learning-based risk stratification tool, called CShock, for patients admitted into the cardiac ICU with acute decompensated heart failure and/or myocardial infarction to predict onset of cardiogenic shock. We prepared a cardiac ICU dataset using MIMIC-III database by annotating with physician adjudicated outcomes. This dataset that consisted of 1500 patients with 204 having cardiogenic/mixed shock was then used to train CShock. The features used to train the model for CShock included patient demographics, cardiac ICU admission diagnoses, routinely measured laboratory values and vital signs, and relevant features manually extracted from echocardiogram and left heart catheterization reports. We externally validated the risk model on the New York University (NYU) Langone Health cardiac ICU database that was also annotated with physician adjudicated outcomes. The external validation cohort consisted of 131 patients with 25 patients experiencing cardiogenic/mixed shock. RESULTS: CShock achieved an area under the receiver operator characteristic curve (AUROC) of 0.821 (95% CI 0.792-0.850). CShock was externally validated in the more contemporary NYU cohort and achieved an AUROC of 0.800 (95% CI 0.717-0.884), demonstrating its generalizability in other cardiac ICUs. Having an elevated heart rate is most predictive of cardiogenic shock development based on Shapley values. The other top ten predictors are having an admission diagnosis of myocardial infarction with ST-segment elevation, having an admission diagnosis of acute decompensated heart failure, Braden Scale, Glasgow Coma Scale, Blood urea nitrogen, Systolic blood pressure, Serum chloride, Serum sodium, and Arterial blood pH. CONCLUSIONS: The novel CShock score has the potential to provide automated detection and early warning for cardiogenic shock and improve the outcomes for the millions of patients who suffer from myocardial infarction and heart failure.

Deciphering RNA splicing logic with interpretable machine learning.

Machine learning methods, particularly neural networks trained on large datasets, are transforming how scientists approach scientific discovery and experimental design. However, current state-of-the-art neural networks are limited by their uninterpretability: Despite their excellent accuracy, they cannot describe how they arrived at their predictions. Here, using an "interpretable-by-design" approach, we present a neural network model that provides insights into RNA splicing, a fundamental process in the transfer of genomic information into functional biochemical products. Although we designed our model to emphasize interpretability, its predictive accuracy is on par with state-of-the-art models. To demonstrate the model's interpretability, we introduce a visualization that, for any given exon, allows us to trace and quantify the entire decision process from input sequence to output splicing prediction. Importantly, the model revealed uncharacterized components of the splicing logic, which we experimentally validated. This study highlights how interpretable machine learning can advance scientific discovery.

DIET: Conditional independence testing with marginal dependence measures of residual information.

Conditional randomization tests (CRTs) assess whether a variable x is predictive of another variable y, having observed covariates z. CRTs require fitting a large number of predictive models, which is often computationally intractable. Existing solutions to reduce the cost of CRTs typically split the dataset into a train and test portion, or rely on heuristics for interactions, both of which lead to a loss in power. We propose the decoupled independence test (DIET), an algorithm that avoids both of these issues by leveraging marginal independence statistics to test conditional independence relationships. DIET tests the marginal independence of two random variables: Fx∣z(x∣z) and Fy∣z(y∣z) where F⋅∣z(⋅∣z) is a conditional cumulative distribution function (CDF) for the distribution p(⋅∣z). These variables are termed "information residuals." We give sufficient conditions for DIET to achieve finite sample type-1 error control and power greater than the type-1 error rate. We then prove that when using the mutual information between the information residuals as a test statistic, DIET yields the most powerful conditionally valid test. Finally, we show DIET achieves higher power than other tractable CRTs on several synthetic and real benchmarks.

Fast kernel-based association testing of non-linear genetic effects for biobank-scale data.

Our knowledge of non-linear genetic effects on complex traits remains limited, in part, due to the modest power to detect such effects. While kernel-based tests offer a versatile approach to test for non-linear relationships between sets of genetic variants and traits, current approaches cannot be applied to Biobank-scale datasets containing hundreds of thousands of individuals. We propose, FastKAST, a kernel-based approach that can test for non-linear effects of a set of variants on a quantitative trait. FastKAST provides calibrated hypothesis tests while enabling analysis of Biobank-scale datasets with hundreds of thousands of unrelated individuals from a homogeneous population. We apply FastKAST to 53 quantitative traits measured across ≈ 300 K unrelated white British individuals in the UK Biobank to detect sets of variants with non-linear effects at genome-wide significance.

Contra: Contrarian statistics for controlled variable selection.

The holdout randomization test (HRT) discovers a set of covariates most predictive of a response. Given the covariate distribution, HRTs can explicitly control the false discovery rate (FDR). However, if this distribution is unknown and must be estimated from data, HRTs can inflate the FDR. To alleviate the inflation of FDR, we propose the contrarian randomization test (CONTRA), which is designed explicitly for scenarios where the covariate distribution must be estimated from data and may even be misspecified. Our key insight is to use an equal mixture of two "contrarian" probabilistic models in determining the importance of a covariate. One model is fit with the real data, while the other is fit using the same data, but with the covariate being tested replaced with samples from an estimate of the covariate distribution. CONTRA is flexible enough to achieve a power of 1 asymptotically, can reduce the FDR compared to state-of-the-art CVS methods when the covariate distribution is misspecified, and is computationally efficient in high dimensions and large sample sizes. We further demonstrate the effectiveness of CONTRA on numerous synthetic benchmarks, and highlight its capabilities on a genetic dataset.

Have We Learned to Explain?: How Interpretability Methods Can Learn to Encode Predictions in their Interpretations.

While the need for interpretable machine learning has been established, many common approaches are slow, lack fidelity, or hard to evaluate. Amortized explanation methods reduce the cost of providing interpretations by learning a global selector model that returns feature importances for a single instance of data. The selector model is trained to optimize the fidelity of the interpretations, as evaluated by a predictor model for the target. Popular methods learn the selector and predictor model in concert, which we show allows predictions to be encoded within interpretations. We introduce EVAL-X as a method to quantitatively evaluate interpretations and REAL-X as an amortized explanation method, which learn a predictor model that approximates the true data generating distribution given any subset of the input. We show EVAL-X can detect when predictions are encoded in interpretations and show the advantages of REAL-X through quantitative and radiologist evaluation.

14-3-3γ prevents centrosome duplication by inhibiting NPM1 function.

14-3-3 proteins bind to ligands via phospho-serine containing consensus motifs. However, the molecular mechanisms underlying complex formation and dissociation between 14-3-3 proteins and their ligands remain unclear. We identified two conserved acidic residues in the 14-3-3 peptide-binding pocket (D129 and E136) that potentially regulate complex formation and dissociation. Altering these residues to alanine led to opposing effects on centrosome duplication. D129A inhibited centrosome duplication, whereas E136A stimulated centrosome amplification. These results were due to the differing abilities of these mutant proteins to form a complex with NPM1. Inhibiting complex formation between NPM1 and 14-3-3γ led to an increase in centrosome duplication and over-rode the ability of D129A to inhibit centrosome duplication. We identify a novel role of 14-3-3γ in regulating centrosome licensing and a novel mechanism underlying the formation and dissociation of 14-3-3 ligand complexes dictated by conserved residues in the 14-3-3 family.

A validated, real-time prediction model for favorable outcomes in hospitalized COVID-19 patients.

The COVID-19 pandemic has challenged front-line clinical decision-making, leading to numerous published prognostic tools. However, few models have been prospectively validated and none report implementation in practice. Here, we use 3345 retrospective and 474 prospective hospitalizations to develop and validate a parsimonious model to identify patients with favorable outcomes within 96 h of a prediction, based on real-time lab values, vital signs, and oxygen support variables. In retrospective and prospective validation, the model achieves high average precision (88.6% 95% CI: [88.4-88.7] and 90.8% [90.8-90.8]) and discrimination (95.1% [95.1-95.2] and 86.8% [86.8-86.9]) respectively. We implemented and integrated the model into the EHR, achieving a positive predictive value of 93.3% with 41% sensitivity. Preliminary results suggest clinicians are adopting these scores into their clinical workflows.

Deep direct likelihood knockoffs.

Predictive modeling often uses black box machine learning methods, such as deep neural networks, to achieve state-of-the-art performance. In scientific domains, the scientist often wishes to discover which features are actually important for making the predictions. These discoveries may lead to costly follow-up experiments and as such it is important that the error rate on discoveries is not too high. Model-X knockoffs [2] enable important features to be discovered with control of the false discovery rate (fdr). However, knockoffs require rich generative models capable of accurately modeling the knockoff features while ensuring they obey the so-called "swap" property. We develop Deep Direct Likelihood Knockoffs (ddlk), which directly minimizes the KL divergence implied by the knockoff swap property. ddlk consists of two stages: it first maximizes the explicit likelihood of the features, then minimizes the KL divergence between the joint distribution of features and knockoffs and any swap between them. To ensure that the generated knockoffs are valid under any possible swap, ddlk uses the Gumbel-Softmax trick to optimize the knockoff generator under the worst-case swap. We find ddlk has higher power than baselines while controlling the false discovery rate on a variety of synthetic and real benchmarks including a task involving a large dataset from one of the epicenters of COVID-19.

Deep dreaming, aberrant salience and psychosis: Connecting the dots by artificial neural networks.

Why some individuals, when presented with unstructured sensory inputs, develop altered perceptions not based in reality, is not well understood. Machine learning approaches can potentially help us understand how the brain normally interprets sensory inputs. Artificial neural networks (ANN) progressively extract higher and higher-level features of sensory input and identify the nature of an object based on a priori information. However, some ANNs which use algorithms such as the "deep-dreaming" developed by Google, allow the network to over-emphasize some objects it "thinks" it recognizes in those areas, and iteratively enhance such outputs leading to representations that appear farther and farther from "reality". We suggest that such "deep dreaming" ANNs may model aberrant salience, a mechanism suggested for pathogenesis of psychosis. Such models can generate testable predictions for psychosis.