Domain adaptation based on rough adjoint inconsistency and optimal transport for identifying autistic patients.

BACKGROUND AND OBJECTIVE: Computer aided diagnosis technology has been widely used to diagnose autism spectrum disorder (ASD) from neural images. The performance of the model usually depends largely on a sufficient number of training samples that reflect the real sample distribution. Due to the lack of labelled neural images data, multisite data are often pooled together to expand the sample size. However, the heterogeneity among sites will inevitably lead to a decline in the generalization of models. To solve this problem, we propose a multisource unsupervised domain adaptation method using rough adjoint inconsistency and optimal transport. METHODS: First, we define the concept of rough adjoint inconsistency and propose a double quantization method based on rough adjoint inconsistency and Dempster-Shafer (D-S) evidence theory to estimate the weight coefficient of each source domain to accurately describe the importance of each source domain to the target domain. Second, using optimal transport theory, we weaken the data distribution differences between domains and solve the problem of class imbalance by adjusting the sampling weights among classes. RESULTS: The ASD recognition accuracy of the proposed method is improved on all eight tasks, which are 70.67%, 64.86%, 62.50%, 70.80%, 73.08%, 71.19%, 75.41% and 75.76%, respectively. Our proposed model achieves superior performance compared to traditional machine learning methods and other recently proposed deep learning model. CONCLUSIONS: Our method demonstrates that the fusion of rough adjoint inconsistency and optimal transport can be a powerful tool for identifying ASD and quantifying the correlations between domains.

Comparative analysis of variable region of white spot syndrome virus genome in Penaeus vannamei in Guangxi, China / 病毒学报

Comparative analysis of variable region ORF14/15 genes of white spot syndrome virus (WSSV) genome in Guangxi Penaeus vannamei (P. vannamei) could provide useful information for the evaluation of genetic diversity and genetic evolutionary relationship among WSSV isolates from Guangxi, China and other places. Based on geographical and temporal considerations, 40 WSSV-positive P. vannamei samples were collected during the period between May 2010 and July 2013 from Beihai, Qinzhou, and Fangchenggang, which were the main P. vannamei production areas in Guangxi, and the variable region ORF14/15 genes of the WSSV genome from all infected samples were amplified by PCR and then subjected to cloning and sequence analysis. Pairwise and multiple alignment analysis was then conducted to evaluate the degree of genetic divergence between different strains. The variable region ORF14/15 genes from 25 of 40 WSSV positive samples were successfully cloned and sequenced; among the ORF14/15 genes of 25 WSSV-positive strains, 22 was 619 bp in length and 3 was 620 bp. All the 25 Guangxi strains carried a 5949-bp deletion in the ORF14/15 region relative to TH-96-II, which has the longest nucleotide sequence in this region; the deletion of Guangxi strains occurred in the middle region of ORF14/15 gene, with only 190 bp and 429 bp/ 430 bp at 5' and 3' ends, respectively, which were coincident with WSSV-IN-05-I in deletion length and position. Sixteen of 25 Guangxi strains had completely identical nucleotide sequences in the variable re gion, and the homology between other strains also exceeded 97.9%. There were single nucleotide substi tution, deletion, and insertion in the ORF14/15 region of Guangxi strains compared with other strains in GenBank. In the phylogenetic tree based on WSSV variable region ORF14/15, the Guangxi strains were closely related and formed a separate branch with Indian strain IN-05-I, but far from other strains in GenBank. The ORF14/15 gene of WSSV isolates in cultured P. vannamei in Guangxi has a large deletion in the middle of the variable region, and the Guangxi WSSV strains show no significant spatio-temporal differences; the Guangxi strains are closer in genetics to Indian strain IN-05-I than other strains in GenBank.