RNA biomarkers enable early and precise disease diagnosis, monitoring, and prognosis, facilitating personalized medicine and targeted therapeutic strategies. However, identification of RNA biomarkers is hindered by the challenge of analyzing relatively small yet high-dimensional transcriptomics datasets, typically comprising fewer than 1000 biospecimens but encompassing hundreds of thousands of RNAs, especially noncoding RNAs. This complexity leads to several limitations in existing methods, such as poor reproducibility on independent datasets, inability to directly process omics data, and difficulty in identifying noncoding RNAs as biomarkers. Additionally, these methods often yield results that lack biological interpretation and clinical utility. To overcome these challenges, we present BAMBI (Biostatistical and Artificial-intelligence Methods for Biomarker Identification), a computational tool integrating biostatistical approaches and machine-learning algorithms. By initially reducing high dimensionality through biologically informed statistical methods followed by machine learning-based feature selection, BAMBI significantly enhances the accuracy and clinical utility of identified RNA biomarkers and also includes noncoding RNA biomarkers that existing methods may overlook. BAMBI outperformed existing methods on both real and simulated datasets by identifying individual and panel biomarkers with fewer RNAs while still ensuring superior prediction accuracy. BAMBI was benchmarked on multiple transcriptomics datasets across diseases, including breast cancer, psoriasis, and leukemia. The prognostic biomarkers for acute myeloid leukemia discovered by BAMBI showed significant correlations with patient survival rates in an independent cohort, highlighting its potential for enhancing clinical outcomes. The software is available on GitHub (https://github.com/CZhouLab/BAMBI).

As technology scales below 32 nm, manufacturers began to integrate both CPU and GPU cores in a single chip, i.e., single-chip heterogeneous processor (SCHP), to improve the throughput of emerging applications. In SCHPs, the CPU and the GPU share the total chip power budget while satisfying their own power constraints, respectively. Consequently, to maximize the overall throughput and/or power efficiency, both power budget and workload should be judiciously allocated to the CPU and the GPU. In this paper, we first demonstrate that optimal allocation of power budget and workload to the CPU and the GPU can provide 13 percent higher throughput than the optimal allocation of workload alone for a single-program workload scenario. Second, we also demonstrate that asymmetric power allocation considering per-program characteristics for a multi-programmed workload scenario can provide 9 percent higher throughput or 24 percent higher power efficiency than the even power allocation per program depending on the optimization objective. Last, we propose effective runtime algorithms that can determine near-optimal or optimal combinations of workload and power budget partitioning for both single- and multi-programmed workload scenarios; the runtime algorithms can achieve 96 and 99 percent of the maximum achievable throughput within 5-8 and 3-5 kernel invocations for single- and multi-programmed workload cases, respectively.