Search for: All records

Abstract Drug screening data from massive bulk gene expression databases can be analyzed to determine the optimal clinical application of cancer drugs. The growing amount of single-cell RNA sequencing (scRNA-seq) data also provides insights into improving therapeutic effectiveness by helping to study the heterogeneity of drug responses for cancer cell subpopulations. Developing computational approaches to predict and interpret cancer drug response in single-cell data collected from clinical samples can be very useful. We propose scDEAL, a deep transfer learning framework for cancer drug response prediction at the single-cell level by integrating large-scale bulk cell-line data. The highlight in scDEAL involves harmonizing drug-related bulk RNA-seq data with scRNA-seq data and transferring the model trained on bulk RNA-seq data to predict drug responses in scRNA-seq. Another feature of scDEAL is the integrated gradient feature interpretation to infer the signature genes of drug resistance mechanisms. We benchmark scDEAL on six scRNA-seq datasets and demonstrate its model interpretability via three case studies focusing on drug response label prediction, gene signature identification, and pseudotime analysis. We believe that scDEAL could help study cell reprogramming, drug selection, and repurposing for improving therapeutic efficacy. 

Abstract—Elastin is a key structural protein and its pathological degradation deterministic in aortic aneurysm (AA) outcomes. Unfortunately, using current diagnostic and clinical surveillance techniques the integrity of the elastic fiber network can only be assessed invasively. To address this, we employed fragmented elastin-targeting gold nanoparticles (EL-AuNPs) as a diagnostic tool for the evaluation of unruptured AAs. Electron dense EL-AuNPs were visualized within AAs using microcomputed tomography (micro-CT) and the corresponding Gold-to-Tissue volume ratios quantified. The Gold-to-Tissue volume ratios correlated strongly with the concentration (0, 0.5, or 10 U/mL) of infused porcine pancreatic elastase and therefore the degree of elastin damage. Hyperspectralmapping confirmed the spatial targeting of the EL-AuNPs to the sites of damaged elastin. Nonparametric Spearman’s rank correlation indicated that the micro-CT-based Gold-to-Tissue volume ratios had a strong correlation with loaded (q = 0.867, p-val = 0.015) and unloaded (q = 0.830, p-val = 0.005) vessel diameter, percent dilation (q = 0.976, p-val = 0.015), circumferential stress (q = 0.673, p-val = 0.007), loaded (q = 2 0.673, p-val = 0.017) and unloaded (q = 2 0.697, p-val = 0.031) wall thicknesses, circumferential stretch (q = 2 0.7234, p-val = 0.018), and lumen area compliance (q = 2 0.831, p-val = 0.003). Likewise, in terms of axial force and axial stress vs. stretch, the post-elastase vessels were stiffer. Collectively, these findings suggest that, when combined with CT imaging, EL-AuNPs can be used as a powerful tool in the non-destructive estimation of mechanical and geometric features of AAs.