Forecasting acute kidney injury and resource utilization in ICU patients using longitudinal, multimodal models.

BACKGROUND: Advances in artificial intelligence (AI) have realized the potential of revolutionizing healthcare, such as predicting disease progression via longitudinal inspection of Electronic Health Records (EHRs) and lab tests from patients admitted to Intensive Care Units (ICU). Although substantial literature exists addressing broad subjects, including the prediction of mortality, length-of-stay, and readmission, studies focusing on forecasting Acute Kidney Injury (AKI), specifically dialysis anticipation like Continuous Renal Replacement Therapy (CRRT) are scarce. The technicality of how to implement AI remains elusive. OBJECTIVE: This study aims to elucidate the important factors and methods that are required to develop effective predictive models of AKI and CRRT for patients admitted to ICU, using EHRs in the Medical Information Mart for Intensive Care (MIMIC) database. METHODS: We conducted a comprehensive comparative analysis of established predictive models, considering both time-series measurements and clinical notes from MIMIC-IV databases. Subsequently, we proposed a novel multi-modal model which integrates embeddings of top-performing unimodal models, including Long Short-Term Memory (LSTM) and BioMedBERT, and leverages both unstructured clinical notes and structured time series measurements derived from EHRs to enable the early prediction of AKI and CRRT. RESULTS: Our multimodal model achieved a lead time of at least 12 h ahead of clinical manifestation, with an Area Under the Receiver Operating Characteristic Curve (AUROC) of 0.888 for AKI and 0.997 for CRRT, as well as an Area Under the Precision Recall Curve (AUPRC) of 0.727 for AKI and 0.840 for CRRT, respectively, which significantly outperformed the baseline models. Additionally, we performed a SHapley Additive exPlanation (SHAP) analysis using the expected gradients algorithm, which highlighted important, previously underappreciated predictive features for AKI and CRRT. CONCLUSION: Our study revealed the importance and the technicality of applying longitudinal, multimodal modeling to improve early prediction of AKI and CRRT, offering insights for timely interventions. The performance and interpretability of our model indicate its potential for further assessment towards clinical applications, to ultimately optimize AKI management and enhance patient outcomes.

Forecasting Acute Kidney Injury and Resource Utilization in ICU patients using longitudinal, multimodal models.

Background: Advances in artificial intelligence (AI) have realized the potential of revolutionizing healthcare, such as predicting disease progression via longitudinal inspection of Electronic Health Records (EHRs) and lab tests from patients admitted to Intensive Care Units (ICU). Although substantial literature exists addressing broad subjects, including the prediction of mortality, length-of-stay, and readmission, studies focusing on forecasting Acute Kidney Injury (AKI), specifically dialysis anticipation like Continuous Renal Replacement Therapy (CRRT) are scarce. The technicality of how to implement AI remains elusive. Objective: This study aims to elucidate the important factors and methods that are required to develop effective predictive models of AKI and CRRT for patients admitted to ICU, using EHRs in the Medical Information Mart for Intensive Care (MIMIC) database. Methods: We conducted a comprehensive comparative analysis of established predictive models, considering both time-series measurements and clinical notes from MIMIC-IV databases. Subsequently, we proposed a novel multi-modal model which integrates embeddings of top-performing unimodal models, including Long Short-Term Memory (LSTM) and BioMedBERT, and leverages both unstructured clinical notes and structured time series measurements derived from EHRs to enable the early prediction of AKI and CRRT. Results: Our multimodal model achieved a lead time of at least 12 hours ahead of clinical manifestation, with an Area Under the Receiver Operating Characteristic Curve (AUROC) of 0.888 for AKI and 0.997 for CRRT, as well as an Area Under the Precision Recall Curve (AUPRC) of 0.727 for AKI and 0.840 for CRRT, respectively, which significantly outperformed the baseline models. Additionally, we performed a SHapley Additive exPlanation (SHAP) analysis using the expected gradients algorithm, which highlighted important, previously underappreciated predictive features for AKI and CRRT. Conclusion: Our study revealed the importance and the technicality of applying longitudinal, multimodal modeling to improve early prediction of AKI and CRRT, offering insights for timely interventions. The performance and interpretability of our model indicate its potential for further assessment towards clinical applications, to ultimately optimize AKI management and enhance patient outcomes.

Elucidating immune-related gene transcriptional programs via factorization of large-scale RNA-profiles.

Recent developments in immunotherapy, including immune checkpoint blockade (ICB) and adoptive cell therapy (ACT), have encountered challenges such as immune-related adverse events and resistance, especially in solid tumors. To advance the field, a deeper understanding of the molecular mechanisms behind treatment responses and resistance is essential. However, the lack of functionally characterized immune-related gene sets has limited data-driven immunological research. To address this gap, we adopted non-negative matrix factorization on 83 human bulk RNA sequencing (RNA-seq) datasets and constructed 28 immune-specific gene sets. After rigorous immunologist-led manual annotations and orthogonal validations across immunological contexts and functional omics data, we demonstrated that these gene sets can be applied to refine pan-cancer immune subtypes, improve ICB response prediction and functionally annotate spatial transcriptomic data. These functional gene sets, informing diverse immune states, will advance our understanding of immunology and cancer research.

Elucidating immune-related gene transcriptional programs via factorization of large-scale RNA-profiles.

Recent developments in immunotherapy, including immune checkpoint blockade (ICB) and adoptive cell therapy, have encountered challenges such as immune-related adverse events and resistance, especially in solid tumors. To advance the field, a deeper understanding of the molecular mechanisms behind treatment responses and resistance is essential. However, the lack of functionally characterized immune-related gene sets has limited data-driven immunological research. To address this gap, we adopted non-negative matrix factorization on 83 human bulk RNA-seq datasets and constructed 28 immune-specific gene sets. After rigorous immunologist-led manual annotations and orthogonal validations across immunological contexts and functional omics data, we demonstrated that these gene sets can be applied to refine pan-cancer immune subtypes, improve ICB response prediction and functionally annotate spatial transcriptomic data. These functional gene sets, informing diverse immune states, will advance our understanding of immunology and cancer research.

Single-nucleotide variant calling in single-cell sequencing data with Monopogen.

Single-cell omics technologies enable molecular characterization of diverse cell types and states, but how the resulting transcriptional and epigenetic profiles depend on the cell's genetic background remains understudied. We describe Monopogen, a computational tool to detect single-nucleotide variants (SNVs) from single-cell sequencing data. Monopogen leverages linkage disequilibrium from external reference panels to identify germline SNVs and detects putative somatic SNVs using allele cosegregating patterns at the cell population level. It can identify 100 K to 3 M germline SNVs achieving a genotyping accuracy of 95%, together with hundreds of putative somatic SNVs. Monopogen-derived genotypes enable global and local ancestry inference and identification of admixed samples. It identifies variants associated with cardiomyocyte metabolic levels and epigenomic programs. It also improves putative somatic SNV detection that enables clonal lineage tracing in primary human clonal hematopoiesis. Monopogen brings together population genetics, cell lineage tracing and single-cell omics to uncover genetic determinants of cellular processes.

PALLAS: Penalized mAximum LikeLihood and pArticle Swarms for Inference of Gene Regulatory Networks From Time Series Data.

We present PALLAS, a practical method for gene regulatory network (GRN) inference from time series data, which employs penalized maximum likelihood and particle swarms for optimization. PALLAS is based on the Partially-Observable Boolean Dynamical System (POBDS) model and thus does not require ad-hoc binarization of the data. The penalty in the likelihood is a LASSO regularization term, which encourages the resulting network to be sparse. PALLAS is able to scale to networks of realistic size under no prior knowledge, by virtue of a novel continuous-discrete Fish School Search particle swarm algorithm for efficient simultaneous maximization of the penalized likelihood over the discrete space of networks and the continuous space of observational parameters. The performance of PALLAS is demonstrated by a comprehensive set of experiments using synthetic data generated from real and artificial networks, as well as real time series microarray and RNA-seq data, where it is compared to several other well-known methods for gene regulatory network inference. The results show that PALLAS can infer GRNs more accurately than other methods, while being capable of working directly on gene expression data, without need of ad-hoc binarization. PALLAS is a fully-fledged program, written in python, and available on GitHub (https://github.com/yukuntan92/PALLAS).

Single-cell CRISPR immune screens reveal immunological roles of tumor intrinsic factors.

Genetic screens are widely exploited to develop novel therapeutic approaches for cancer treatment. With recent advances in single-cell technology, single-cell CRISPR screen (scCRISPR) platforms provide opportunities for target validation and mechanistic studies in a high-throughput manner. Here, we aim to establish scCRISPR platforms which are suitable for immune-related screens involving multiple cell types. We integrated two scCRISPR platforms, namely Perturb-seq and CROP-seq, with both in vitro and in vivo immune screens. By leveraging previously generated resources, we optimized experimental conditions and data analysis pipelines to achieve better consistency between results from high-throughput and individual validations. Furthermore, we evaluated the performance of scCRISPR immune screens in determining underlying mechanisms of tumor intrinsic immune regulation. Our results showed that scCRISPR platforms can simultaneously characterize gene expression profiles and perturbation effects present in individual cells in different immune screen conditions. Results from scCRISPR immune screens also predict transcriptional phenotype associated with clinical responses to cancer immunotherapy. More importantly, scCRISPR screen platforms reveal the interactive relationship between targeting tumor intrinsic factors and T cell-mediated antitumor immune response which cannot be easily assessed by bulk RNA-seq. Collectively, scCRISPR immune screens provide scalable and reliable platforms to elucidate molecular determinants of tumor immune resistance.

Spatially resolved transcriptomics of high-grade serous ovarian carcinoma.

Bulk and single-cell RNA sequencing do not provide full characterization of tissue spatial diversity in cancer samples, and currently available in situ techniques (multiplex immunohistochemistry and imaging mass cytometry) allow for only limited analysis of a small number of targets. The current study represents the first comprehensive approach to spatial transcriptomics of high-grade serous ovarian carcinoma using intact tumor tissue. We selected a small cohort of patients with highly annotated high-grade serous ovarian carcinoma, categorized them by response to neoadjuvant chemotherapy (poor or excellent), and analyzed pre-treatment tumor tissue specimens. Our study uncovered extensive differences in tumor composition between the poor responders and excellent responders to chemotherapy, related to cell cluster organization and localization. This in-depth characterization of high-grade serous ovarian carcinoma tumor tissue from poor and excellent responders showed that spatial interactions between cell clusters may influence chemo-responsiveness more than cluster composition alone.

Molecular Correlates of Venous Thromboembolism (VTE) in Ovarian Cancer.

BACKGROUND: The incidence of venous thromboembolism (VTE) in patients with ovarian cancer is higher than most solid tumors, ranging between 10-30%, and a diagnosis of VTE in this patient population is associated with worse oncologic outcomes. The tumor-specific molecular factors that may lead to the development of VTE are not well understood. OBJECTIVES: The aim of this study was to identify molecular features present in ovarian tumors of patients with VTE compared to those without. METHODS: We performed a multiplatform omics analysis incorporating RNA and DNA sequencing, quantitative proteomics, as well as immune cell profiling of high-grade serous ovarian carcinoma (HGSC) samples from a cohort of 32 patients with or without VTE. RESULTS: Pathway analyses revealed upregulation of both inflammatory and coagulation pathways in the VTE group. While DNA whole-exome sequencing failed to identify significant coding alterations between the groups, the results of an integrated proteomic and RNA sequencing analysis indicated that there is a relationship between VTE and the expression of platelet-derived growth factor subunit B (PDGFB) and extracellular proteins in tumor cells, namely collagens, that are correlated with the formation of thrombosis. CONCLUSIONS: In this comprehensive analysis of HGSC tumor tissues from patients with and without VTE, we identified markers unique to the VTE group that could contribute to development of thrombosis. Our findings provide additional insights into the molecular alterations underlying the development of VTE in ovarian cancer patients and invite further investigation into potential predictive biomarkers of VTE in ovarian cancer.

A stochastic metapopulation state-space approach to modeling and estimating COVID-19 spread.

Mathematical models are widely recognized as an important tool for analyzing and understanding the dynamics of infectious disease outbreaks, predict their future trends, and evaluate public health intervention measures for disease control and elimination. We propose a novel stochastic metapopulation state-space model for COVID-19 transmission, which is based on a discrete-time spatio-temporal susceptible, exposed, infected, recovered, and deceased (SEIRD) model. The proposed framework allows the hidden SEIRD states and unknown transmission parameters to be estimated from noisy, incomplete time series of reported epidemiological data, by application of unscented Kalman filtering (UKF), maximum-likelihood adaptive filtering, and metaheuristic optimization. Experiments using both synthetic data and real data from the Fall 2020 COVID-19 wave in the state of Texas demonstrate the effectiveness of the proposed model.