Background Primary immunodeficiency diseases represent an expanding set of heterogeneous conditions which are difficult to recognize clinically. Diagnostic rates outside of the newborn period have not changed appreciably. This concern underscores a need for novel methods of disease detection. Objective We built a Bayesian network to provide real-time risk assessment about primary immunodeficiency and to facilitate prescriptive analytics for initiating the most appropriate diagnostic work up. Our goal is to improve diagnostic rates for primary immunodeficiency and shorten time to diagnosis. We aimed to use readily available health record data and a small training dataset to prove utility in diagnosing patients with relatively rare features. Methods We extracted data from the Texas Children’s Hospital electronic health record on a large population of primary immunodeficiency patients (n = 1762) and appropriately-matched set of controls (n = 1698). From the cohorts, clinically relevant prior probabilities were calculated enabling construction of a Bayesian network probabilistic model(PI Prob). Our model was constructed with clinical-immunology domain expertise, trained on a balanced cohort of 100 cases-controls and validated on an unseen balanced cohort of 150 cases-controls. Performance was measured by area under the receiver operator characteristic curve (AUROC). We also compared our network performance to classic machine learning model performance on the same dataset. Results PI Prob was accurate in classifying immunodeficiency patients from controls (AUROC = 0.945; p<0.0001) at a risk threshold of ≥6%. Additionally, the model was 89% accurate for categorizing validation cohort members into appropriate International Union of Immunological Societies diagnostic categories. Our network outperformed 3 other machine learning models and provides superior transparency with a prescriptive output element. Conclusion Artificial intelligence methods can classify risk for primary immunodeficiency and guide management. PI Prob enables accurate, objective decision making about risk and guides the user towards the appropriate diagnostic evaluation for patients with recurrent infections. Probabilistic models can be trained with small datasets underscoring their utility for rare disease detection given appropriate domain expertise for feature selection and network construction.

Cell-cell communication events (CEs) mediated by multiple ligand-receptor pairs construct a complex intercellular signaling network. Usually only a subset of CEs directly works for a specific downstream response in certain microenvironments. We call them functional communication events (FCEs). Spatial transcriptomic methods can profile the spatial distribution of gene expression levels of ligands, receptors, and their downstream genes. This provides a new possibility for revealing the holographic network of cell-cell communication. We developed HoloNet, a computational method for decoding FCEs using spatial transcriptomic data. We modeled CEs as a multi-view network, developed an attention-based graph learning model on the network to predict the target gene expression, and decoded the FCEs for specific downstream genes by interpreting the trained model. We applied HoloNet on two breast cancer Visium datasets to reveal the communication landscapes in breast cancer microenvironments. It detected ligand-receptor signals triggering the expression changes of invasion-related genes in stromal cells surrounding tumors. The experiments showed that HoloNet is a powerful tool on spatial transcriptomic data to help understand the shaping of cellular phenotypes through cell-cell communication in a microenvironment.

No abstract

No abstract

scCRISPR-seq is an emerging high-throughput CRISPR screening technology that combines CIRPSR screening with single-cell sequencing technologies. It provides rich information on gene regulation. When performing scCRISPR-seq in a population of heterogeneous cells, the observed cellular response in perturbed cells may be caused not only by the perturbation, but also by the infection bias of guide RNAs (gRNAs) mainly contributed by intrinsic differences of cell clusters. The mixing of these effects poisons gene regulation studies. We developed scDecouple to decouple the true cellular response of the perturbation from the influence of infection bias. It models the distribution of perturbed cells and iteratively finds the maximum likelihood of cell cluster proportions as well as the real cellular response for each gRNA. We demonstrated its performance on a series of simulation experiments. By applying scDecouple to real CROP-seq data, we found that scDecouple could enhance biological discovery by detecting perturbation-related genes more critically. It helps to better study gene function and identify disease targets via scCRISPR-seq, especially with heterogeneous samples or complex gRNA libraries.