ABSTRACT
Memristive computing system (MCS), with the feature of in-memory computing capability, for artificial neural networks (ANNs) deployment showing low power and massive parallelism, is a promising alternative for traditional Von-Neumann architecture computing system. However, because of the various non-idealities of both peripheral circuits and memristor array, the performance of the practical MCS tends to be significantly reduced. In this work, a linear compensation method (LCM) is proposed for the performance improvement of MCS under the effect of non-idealities. By considering the effects of various non-ideal states in the MCS as a whole, the output error of the MCS under different conditions is investigated. Then, a mathematic model for the output error is established based on the experimental data. Furthermore, the MCS is researched at the physical circuit level as well, in order to analyze the specific way in which the non-idealities affect the output current. Finally, based on the established mathematical model, the LCM output current is compensated in real time to improve the system performance. The effectiveness of LCM is verified and showing outstanding performance in the residual neural network-34 network architecture, which is easily affected by the non-idealities in hardware. The proposed LCM can be naturally integrated into the operation processes of MCS, paving the way for optimizing the deployment on generic ANN hardware based on the memristor.
ABSTRACT
In this study, differences in the protein content and functional and physicochemical properties of four varieties of egg white (EW) were studied by adding 4-10% sucrose or NaCl and then heating them at 70 °C for 3 min. According to a high-performance liquid chromatography (HPLC) analysis, the percentages of ovalbumin, lysozyme and ovotransferrin rose with an increase in the NaCl or sucrose concentration; however, the percentages of ovomucin and ovomucoid decreased. Furthermore, the foaming properties, gel properties, particle size, α-helixes, ß-sheets, sulfhydryl groups and disulfide bond content also increased, whereas the content of ß-turns and random coils decreased. In addition, the total soluble protein content and functional and physicochemical properties of black bone (BB) chicken and Gu-shi (GS) EWs were higher than those of Hy-Line brown (HY-LINE) and Harbin White (HW) Ews (p < 0.05). Subsequently, transmission electron microscopy (TEM) confirmed the changes in the EW protein structure in the four varieties of Ews. As the aggregations increased, the functional and physicochemical properties decreased. The protein content and functional and physicochemical properties of Ews after heating were correlated with the concentration of NaCl and sucrose and the EW varieties.
ABSTRACT
The clinical applicability of the whole-exome sequencing (WES) in estimating tumor mutational burden (TMB) is currently limited by high cost, time-consuming and tissue availability. And given to the differences in the mutational landscapes among different types of cancer, we aimed to develop a cancer-specific signature to estimate TMB for right-sided colon cancer patients (RCC). Using WES data of 315 RCC patients, we identified the exons in which the number of mutational sites of the coding DNA sequences associated with TMB through linear regression analysis. Then, among these exons, we extracted a signature composed by 102 exons (~0.13 Mbp) through a heuristic selection procedure. The TMB estimated by the signature was highly correlated with those calculated by WES in the discovery dataset (R2=0.9869) and three independent validation datasets (R2=0.9351, R2=0.8063 and R2=0.9527, respectively). And the performance of the signature was superior to a colorectal-specific TMB estimation model contained 22 genes (~0.24 Mbp). Moreover, between TMB-high and TMB-low RCC patients, there were significantly differences in the frequencies of microsatellite instability status, CpG island methylator phenotype, BRAF, KRAS and POLE/POLD1 mutation status (p<0.01). However, the performances of the signature in other types of cancer were dramatically degraded (left-sided colon cancer, R2=0.7849 and 0.9407, respectively; rectum, R2=0.5955 and R2=0.965, respectively; breast cancer, R2=0.8444; lung cancer, R2=0.5963), suggesting that it was necessary to develop cancer-specific TMB estimated signatures to estimate precisely the TMB in different types of cancer. In summary, we developed an exon signature that can accurately estimate TMB in RCC patients, and the cost and time required for the assessment of TMB can be considerably decreased, making it more suitable for blood and/or biopsy samples.