Patrick McKeever

We present an interactive Data Integrator to facilitate uniform data processing and integration with datasets generated by the NIH funded MorPhiC (Molecular Phenotypes of Null Alleles in Cells) program. In the dataset identification step, users can select published bulk RNA sequencing (RNA-seq) datasets generated by the MorPhiC program, specify external datasets from the Gene Expression Omnibus (GEO) repository, or upload raw RNA-seq data. Subsequently, users can select a subset of samples and indicate experimental design by assigning groups to samples. External data will then be processed using the MorPhiC approved bulk RNA-seq analytical pipeline, and differential expression will be inferred. Users also have the option to download the uniformly processed counts table.

Single-cell RNA sequencing (scRNA-seq) data is subject to strict access control due to its sensitive nature, motivating the use of synthetic data generation (SDG) for privacy-preserving data sharing. We present the first adversarial privacy attack that performs meaningfully above random guessing against state-of-the-art scRNA-seq SDG methods. Our attack enables donor-level membership inference, demonstrating that leading SDG techniques fail to adequately mask which individuals were used to train the generator. We show that privacy leakage increases as the number of training donors decreases. Although the attack is designed to exploit vulnerabilities in scDesign2, we find that it also succeeds against synthetic data generated by other leading methods, including scDesign3 and scVI. This transferability indicates that an adversary can infer sensitive information from synthetic data without access to the training procedure, model parameters, or even the underlying generation algorithm. Finally, we investigate the use of perturbation with noise during the SDG process as a first-line defense, empirically evaluating its effectiveness in neutralizing the attack and its impact on utility.

The exponential growth of omics data requires novel strategies for storage, transfer, and processing of said data. We present a scheduler based on the Temporal.io workflow framework which enables two key optimizations of bioinformatics workflows. Firstly, we enable users to transparently map workflow steps to diverse execution environments, including high-performance computing (HPC) resources managed by the SLURM resource manager through an easy-to-use graphical user interface. Secondly, we enable asynchronous execution of workflows, a feature which guarantees that workflows will achieve reasonable resource utilization even when the scheduler cannot make use of a system’s full RAM and CPU resources. Thirdly, we propose a universal, platform agnostic JSON representation of workflows that allows platform-specific execution details to be abstracted away from the core scientific logic. Our work includes a custom executor plugin that supports translation of workflows from an external language, such as Nextflow, to our universal JSON format. Finally, we develop a graphical user interface to make our scheduler easy-to-use for non-technical users. When benchmarked on a bulk RNA sequencing workflow, these features reduced the cost and time requirements. We illustrated the merits of our cross-platform method using credit allocations from federally funded supercomputers.