RESUMO
The COVID-19 medical diagnosis method based on individual's chest X-ray (CXR) is achieved difficultly in the initial research, owing to difficulties in identifying CXR data of COVID-19 individuals. At the beginning of the study, infected individuals' CXRs were scarce. The combination of artificial intelligence and medical diagnosis has been advanced and popular. To solve the difficulties, the interpretability analysis of AI model was used to explore the pathological characteristics of CXR samples infected with COVID-19 and assist medical diagnosis. The dataset was expanded by data augmentation to avoid overfitting. Transfer learning was used to test different pre-trained models and the unique output layers were designed to complete the model training with few samples. In this study, the output results of four pre-trained models were compared in three different output layers, and the results after data augmentation were compared with the results of the original dataset. The control variable method was used to conduct independent tests of 24 groups. Finally, 99.23% accuracy and 98% recall rate were obtained, and the visual results of CXR interpretability analysis were displayed. The network of COVID-19 interpretable diagnosis algorithm has the characteristics of high generalization and lightweight. It can be quickly applied to other urgent tasks with insufficient experimental data. At the same time, interpretability analysis brings new possibilities for medical diagnosis.
RESUMO
We have developed sparsely cross-linked "nanogels", subcolloidal polymer structures composed of covalently linked, linear polyacrylamide chains, as novel replaceable DNA sequencing matrixes for capillary and microchip electrophoresis. Nanogels were synthesized via inverse emulsion (water-in-oil) copolymerization of acrylamide and a low percentage (approximately 10(-4) mol %) of N,N-methylene bisacrylamide (Bis). Nanogels and nanogel networks were characterized by multiangle laser light scattering and rheometry, respectively, and tested for DNA sequencing in both capillaries and chips with four-color LIF detection. Typical nanogels have an average radius of approximately 230 nm, with approximately 75% of chains incorporating a Bis cross-linker. The properties and performance of nanogel matrixes are compared here to those of a linear polyacrylamide (LPA) network, matched for both polymer weight-average molar mass (M(w)) and the extent of interchain entanglements (c/c). At sequencing concentrations, the two matrixes have similar flow characteristics, important for capillary and microchip loading. However, because of the physical network stability provided by the internally cross-linked structure of the nanogels, substantially longer average read lengths are obtained under standard conditions with the nanogel matrix at a 98.5% accuracy of base-calling (for CE: 680 bases, an 18.7% improvement over LPA, with the best reads as long as 726 bases, compared to 568 bases for the LPA matrix). We further investigated the use of the nanogel matrixes in a high-throughput microfabricated DNA sequencing device consists of 96 separation channels densely fabricated on a 6-in. glass wafer. Again, preliminary DNA sequencing results show that the nanogel matrixes are capable of delivering significantly longer average read length, compared to an LPA matrix of comparable properties. Moreover, nanogel matrixes require 30% less polymer per unit volume than LPA. The addition of a small amount of low molar mass LPA or ultrahigh molar mass LPA to the optimized nanogel sequencing matrix further improves read length as well as the reproducibility of read length (RSD < 1.6%). This is the first report of a replaceable DNA sequencing matrix that provides better performance than LPA, in a side-by-side comparison of polymer matrixes appropriately matched for molar mass and the extent of interchain entanglements. These results could have significant implications for the improvement of microchip-based DNA sequencing technology.
