Capping-layer-mediated lattice mismatch and redox reaction in SrTiO3-based bilayers.

It is well known that the traditional two-dimensional electron system (2DES) hosted by the SrTiO3substrate can exhibit diverse electronic states by modifying the capping layer in heterostructures. However, such capping layer engineering is less studied in the SrTiO3-layer-carried 2DES (or bilayer 2DES), which is different from the traditional one on transport properties but more applicable to the thin-film devices. Here, several SrTiO3bilayers are fabricated by growing various crystalline and amorphous oxide capping layers on the epitaxial SrTiO3layers. For the crystalline bilayer 2DES, the monotonical reduction on the interfacial conductance, as well as carrier mobility, is recorded on increasing the lattice mismatch between the capping layers and epitaxial SrTiO3layer. The mobility edge raised by the interfacial disorders is highlighted in the crystalline bilayer 2DES. On the other hand, when increasing the concentration of Al with high oxygen affinity in the capping layer, the amorphous bilayer 2DES becomes more conductive accompanied by the enhanced carrier mobility but almost constant carrier density. This observation cannot be explained by the simple redox-reaction model, and the interfacial charge screening and band bending need to be considered. Moreover, when the capping oxide layers have the same chemical composition but with different forms, the crystalline 2DES with a large lattice mismatch is more insulating than its amorphous counterpart, and vice versa. Our results shed some light on understanding the different dominant role in forming the bilayer 2DES using crystalline and amorphous oxide capping layer, which may be applicable in designing other functional oxide interfaces.

Genetic Mechanism Revealed of Age-Related Macular Degeneration Based on Fusion of Statistics and Machine Learning Method.

Age-related macular degeneration (AMD) is the most common cause of irreversible vision loss in the developed world which affects the quality of life for millions of elderly individuals worldwide. Genome-wide association studies (GWAS) have identified genetic variants at 34 loci contributing to AMD. To better understand the disease pathogenesis and identify causal genes for AMD, we applied random walk (RW) and support vector machine (SVM) to identify AMD-related genes based on gene interaction relationship and significance of genes. Our model achieved 0.927 of area under the curve (AUC), and 65 novel genes have been identified as AMD-related genes. To verify our results, a statistics method called summary data-based Mendelian randomization (SMR) has been implemented to integrate GWAS data and transcriptome data to verify AMD susceptibility-related genes. We found 45 genes are related to AMD by SMR. Among these genes, 37 genes overlap with those found by SVM-RW. Finally, we revealed the biological process of genetic mutations leading to changes in gene expression leading to AMD. Our results reveal the genetic pathogenic factors and related mechanisms of AMD.