Advancements in sequencing have led to the proliferation of multi-omic profiles of human cells under different conditions and perturbations. In addition, several databases have amassed information about pathways and gene “signatures” – patterns of gene expression associated with specific cellular and phenotypic contexts. An important current challenge in systems biology is to leverage such knowledge about gene coordination to maximize the predictive power and generalization of models applied to high-throughput datasets. However, few such integrative approaches exist that also provide interpretable results quantifying the importance of individual genes and pathways to model accuracy. We introduce AKLI-MATE, a first kernel-based stacked learner that seamlessly incorporates multi-omics feature data with prior information in the form of pathways for either regression or classification tasks. AKLIMATE uses a novel multiple-kernel learning framework where individual kernels capture the prediction propensities recorded in random forests, each built from a specific pathway gene set that integrates all omics data for its member genes. AKLIMATE outperforms state-of-the-art methods on diverse phenotype learning tasks, including predicting microsatellite instability in endometrial and colorectal cancer, survival in breast cancer, and cell line response to gene knockdowns. We show how AKLIMATE is able to connect feature data across data platforms through their common pathways to identify examples of several known and novel contributors of cancer and synthetic lethality.

BACKGROUND: Many national initiatives focus on promoting early hospital arrival of patients with acute ischemic stroke (AIS) because treatment effectiveness is time-dependent. However, several studies reported time-delays in hospital arrival, especially during the COVID-19 pandemic. Our purpose was to evaluate the 10-year trends in last known well to arrival (LKWA) time and assess disparities in patients with AIS. METHODS: A retrospective study of consecutive patients with AIS in the United States admitted to a large, socioeconomically diverse health care system in the New York metropolitan area was performed from 2014 to 2023. LKWA time groups were categorized according to treatment eligibility: 0 to 4.5, >4.5 to 24, and >24 hours. Demographic and clinical characteristics, treatment utilization, and modified Rankin Scale at discharge were extracted from electronic health records. Trend, bivariable, and multivariable logistic regression analyses were conducted. RESULTS: A total of 11 563 patients with AIS were included with 53% (6163) LKWA 0 to 4.5, 34% (3988) LKWA >4.5 to 24, and 12% (1412) LKWA >24 groups. From 2014 to 2023, there was a significant downtrend in the early LKWA 0 to 4.5 (61%–46%) with uptrends in the later LKWA >4.5 to 24 (31%–43%) and LKWA >24 (8%–11%) groups ( P 4.5 groups, the gap widened between racial categories after COVID (2021–2023; P =0.004). Compared with LKWA 0 to 4.5, the LKWA >4.5 to 24 group was less likely to receive endovascular thrombectomy ( P <0.001) and more likely to have worse outcomes (modified Rankin Scale, 2–5; P 4.5 groups were more likely to be older >80 years of age (odds ratio, 1.33 [95% CIs, 1.11–1.58]), men (1.11 [1.03–1.20]), Black patients (1.21 [1.09–1.34]), Asian patients (1.20 [1.03–1.39]), Medicaid insurance (1.18 [1.08–1.29]), and low-income <$80 000 (1.39 [1.20–1.61]). CONCLUSIONS: In the past decade, there was a significant uptrend in patients with AIS arriving in the late LKWA >4.5 groups. Socioeconomic disparities were observed with a persistent uptrend in non-White patients in the late LKWA >4.5 groups after the COVID pandemic. These findings highlight the need to implement targeted efforts to improve disparities in LKWA time in patients with AIS.

Stroke is a leading cause of death and foremost cause of disability in the United States, and extreme heat is the leading weather-related cause of death. We examined the association between ambient temperature and acute ischemic stroke (AIS) admissions in New York State. We leveraged an existing cohort of consecutive AIS admissions presenting to a large healthcare system during a ten-year period (January 2015–December 2024), identified based on discharge diagnosis codes for ischemic stroke. Meteorological data from gridMET (~4 km grids) were used to produce daily averages of maximum temperature (Tmax) for ZIP code tabulation area (ZCTA) boundaries in the New York metropolitan area, linked to residential ZIP codes. Time-stratified case-crossover models with distributed lags (0–1 days) were used to estimate the association between daily maximum temperature (evaluated as a 10 °F increase, 95th percentile temperature vs. median temperature, and 95 °F vs. median temperature) and AIS admissions, both for the entire year and restricted to warm season months of May–October. There were 27,926 AIS admissions from 2015 to 2024. The mean (standard deviation, SD) age was 71.7 (14.4) years, and 13,934 (49.9%) were female, while the mean (SD) Tmax was 62.8 (17.3) °F, and the mean 95th percentile Tmax was 87.9 °F. The relative risk (95% confidence interval) of AIS admissions was strongest on the same day of admission (10 °F increase: 1.041 (1.019, 1.064); 95th percentile: 1.105 (1.047, 1.165); 95 °F: 1.132 (1.059, 1.210), and effects were similar in warm season months. High daily maximum temperatures were associated with an increased risk of AIS admissions, with the highest risk on the same day of admission. These findings support the short-term impact of temperature on AIS admissions and can inform future public health efforts to reduce heat-related AIS risk in affected populations.