The growing number of mobile devices used by terrestrial users has significantly amplified the traffic load on cellular networks. Especially in urban environments, the high traffic demand brought about by dense user populations has bottlenecked network resources. The Space-Air-Ground-Integrated Network (SAGIN) provides a new solution to cope with this demand, enhancing data transmission efficiency through a multi-layered network structure. However, the heterogeneous and dynamic nature of SAGIN also poses significant management and resource allocation challenges. In this paper, we propose a cloud-edge-SAGIN framework for multi-UAV assisted traffic offloading based on Hierarchical Federated Learning (HFL), aiming to improve the traffic offloading ratio while optimizing the offloading resource allocation. HFL is used instead of traditional Federated Learning (FL) to solve problems such as irrational resource allocation due to heterogeneity in SAGIN. Specifically, the framework applies a hierarchical federated average algorithm and sets a reward function at the ground level, aiming to obtain better model parameters, improve model accuracy at aggregation, enhance UAV traffic offloading ratio, and optimize its scheduling and resource allocation. In addition, an improved Reinforcement Learning (RL) algorithm TD3-A4C is designed in this paper to assist UAVs in realizing intelligent decision-making, reducing communication latency, and further improving resource utilization efficiency. Simulation results demonstrate that the proposed framework and algorithms display superior performance across all dimensions and offer robust support for the comprehensive investigation of intelligent traffic offloading networks.

With the exponential growth in mobile data volume, cellular networks are under severe capacity pressure. To address this issue, Unmanned Aerial Vehicles (UAVs) are being used as mobile Base Stations (BSs) for traffic offloading. However, coordinating and scheduling traffic across multiple UAVs and BSs remains a challenge in complex environments. This paper proposes a solution that optimizes UAV deployment locations and user resource allocation, the goal is to maximize traffic offloading and minimize UAV energy consumption simultaneously. We introduce a hierarchical intelligent traffic offloading network optimization framework based on Deep Federated Learning (DFL). Through federated learning, the UAV swarm is organized hierarchically. Additionally, we developed the CPRAFT algorithm, which uses capacity values as criterion to select the Leader UAV (L-UAV). The L-UAV then becomes the top-level central server for model aggregation in the federated learning environment. Furthermore, we formalize the traffic offloading problem as a Markov Decision Process (MDP). Based on MDP, this paper proposes FL-SNTD3 algorithm to optimize dynamic decision-making, which adapts to the ever-changing network environment and fluctuating traffic demands. Simulation experiments demonstrate that the proposed framework and algorithm exhibit outstanding performance in various aspects, providing robust support for future research in intelligent traffic offloading networks.

Each patient's cancer consists of multiple cell subpopulations that are inherently heterogeneous and may develop differing phenotypes such as drug sensitivity or resistance. A personalized treatment regimen should therefore target multiple oncoproteins in the cancer cell populations that are driving the treatment resistance or disease progression in a given patient to provide maximal therapeutic effect, while avoiding severe co-inhibition of non-malignant cells that would lead to toxic side effects. To address the intra- and inter-tumoral heterogeneity when designing combinatorial treatment regimens for cancer patients, we have implemented a machine learning-based platform to guide identification of safe and effective combinatorial treatments that selectively inhibit cancer-related dysfunctions or resistance mechanisms in individual patients. In this case study, we show how the platform enables prediction of cancer-selective drug combinations for patients with high-grade serous ovarian cancer using single-cell imaging cytometry drug response assay, combined with genome-wide transcriptomic and genetic profiles. The platform makes use of drug-target interaction networks to prioritize those combinations that warrant further preclinical testing in scarce patient-derived primary cells. During the case study in ovarian cancer patients, we investigated (i) the relative performance of various ensemble learning algorithms for drug response prediction, (ii) the use of matched single-cell RNA-sequencing data to deconvolute cell population-specific transcriptome profiles from bulk RNA-seq data, (iii) and whether multi-patient or patient-specific predictive models lead to better predictive accuracy. The general platform and the comparison results are expected to become useful for future studies that use similar predictive approaches also in other cancer types.

MOTIVATION: A major challenge in analyzing cancer patient transcriptomes is that the tumors are inherently heterogeneous and evolving. We analyzed 214 bulk RNA samples of a longitudinal, prospective ovarian cancer cohort and found that the sample composition changes systematically due to chemotherapy and between the anatomical sites, preventing direct comparison of treatment-naive and treated samples. RESULTS: To overcome this, we developed PRISM, a latent statistical framework to simultaneously extract the sample composition and cell-type-specific whole-transcriptome profiles adapted to each individual sample. Our results indicate that the PRISM-derived composition-free transcriptomic profiles and signatures derived from them predict the patient response better than the composite raw bulk data. We validated our findings in independent ovarian cancer and melanoma cohorts, and verified that PRISM accurately estimates the composition and cell-type-specific expression through whole-genome sequencing and RNA in situ hybridization experiments. AVAILABILITYAND IMPLEMENTATION: https://bitbucket.org/anthakki/prism. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Single-cell RNA-sequencing (scRNA-seq) has made it possible to profile gene expression in tissues at high resolution. An important preprocessing step prior to performing downstream analyses is to identify and remove cells with poor or degraded sample quality using quality control (QC) metrics. Two widely used QC metrics to identify a 'low-quality' cell are (i) if the cell includes a high proportion of reads that map to mitochondrial DNA (mtDNA) encoded genes and (ii) if a small number of genes are detected. Current best practices use these QC metrics independently with either arbitrary, uniform thresholds (e.g. 5%) or biological context-dependent (e.g. species) thresholds, and fail to jointly model these metrics in a data-driven manner. Current practices are often overly stringent and especially untenable on certain types of tissues, such as archived tumor tissues, or tissues associated with mitochondrial function, such as kidney tissue [1]. We propose a data-driven QC metric (miQC) that jointly models both the proportion of reads mapping to mtDNA genes and the number of detected genes with mixture models in a probabilistic framework to predict the low-quality cells in a given dataset. We demonstrate how our QC metric easily adapts to different types of single-cell datasets to remove low-quality cells while preserving high-quality cells that can be used for downstream analyses. Our software package is available at https://bioconductor.org/packages/miQC.

SUMMARY: Anduril is an analysis and integration framework that facilitates the design, use, parallelization and reproducibility of bioinformatics workflows. Anduril has been upgraded to use Scala for pipeline construction, which simplifies software maintenance, and facilitates design of complex pipelines. Additionally, Anduril's bioinformatics repository has been expanded with multiple components, and tutorial pipelines, for next-generation sequencing data analysis. AVAILABILITYAND IMPLEMENTATION: Freely available at http://anduril.org. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Motivation: DNA methylation aberrations are common in many cancer types. A major challenge hindering comparison of patient-derived samples is that they comprise of heterogeneous collection of cancer and microenvironment cells. We present a computational method that allows comparing cancer methylomes in two or more heterogeneous tumor samples featuring differing, unknown fraction of cancer cells. The method is unique in that it allows comparison also in the absence of normal cell control samples and without prior tumor purity estimates, as these are often unavailable or unreliable in clinical samples. Results: We use simulations and next-generation methylome, RNA and whole-genome sequencing data from two cancer types to demonstrate that the method is accurate and outperforms alternatives. The results show that our method adapts well to various cancer types and to a wide range of tumor content, and works robustly without a control or with controls derived from various sources. Availability and implementation: The method is freely available at https://bitbucket.org/anthakki/dmml. Supplementary information: Supplementary data are available at Bioinformatics online.