CLIP-Driven Prototype Network for Few-Shot Semantic Segmentation.

Recent research has shown that visual-text pretrained models perform well in traditional vision tasks. CLIP, as the most influential work, has garnered significant attention from researchers. Thanks to its excellent visual representation capabilities, many recent studies have used CLIP for pixel-level tasks. We explore the potential abilities of CLIP in the field of few-shot segmentation. The current mainstream approach is to utilize support and query features to generate class prototypes and then use the prototype features to match image features. We propose a new method that utilizes CLIP to extract text features for a specific class. These text features are then used as training samples to participate in the model's training process. The addition of text features enables model to extract features that contain richer semantic information, thus making it easier to capture potential class information. To better match the query image features, we also propose a new prototype generation method that incorporates multi-modal fusion features of text and images in the prototype generation process. Adaptive query prototypes were generated by combining foreground and background information from the images with the multi-modal support prototype, thereby allowing for a better matching of image features and improved segmentation accuracy. We provide a new perspective to the task of few-shot segmentation in multi-modal scenarios. Experiments demonstrate that our proposed method achieves excellent results on two common datasets, PASCAL-5i and COCO-20i.

Prognostic implications of the EGFR polymorphism rs763317 and clinical variables among young Chinese lung cancer population.

The 5-year survival rate for patients with lung cancer, the world's second most frequent malignant tumor, is less than 20%, and its prognosis cannot be clearly predicted. Our aim was to analyze the epidermal growth factor receptor (EGFR) rs763317 (G>A) single nucleotide polymorphism and its association with prognosis in Chinese Han lung cancer patients. 839 patients with primary lung cancer were recruited, and genomic DNA was extracted and genotyped by SNPscan. Kaplan-Meier technique and multivariate Cox proportional hazards model were used to analyze the association between prognosis and EGFR polymorphism rs763317. A significant association after stratification by age, significantly increased lung cancer risk was associated with the AA homozygous genotype of rs763317 (adjusted hazard ratio = 2.53, 95% CI: 1.31-4.88, p=0.005), and conferred a poor survival for lung cancer patients (MST: median survival time: 13.6 months) compared with GG genotype (MST: 41.5 months), and in the recessive model AA genotype (AA vs. GG + GA; adjusted hazard ratio = 2.57, 95% CI: 1.34-4.93, p=0.004) who were young (<60 years) had a significantly increased risk of death. The EGFR polymorphism rs763617 might serve as a significant genetic marker for predicting the prognosis of lung cancer.

Positional cloning and next-generation sequencing identified a TGM6 mutation in a large Chinese pedigree with acute myeloid leukaemia.

An inherited predisposition to acute myeloid leukaemia (AML) is exceedingly rare, but the investigation of these families will aid in the delineation of the underlying mechanisms of the more common, sporadic cases. Three AML predisposition genes, RUNX1, CEBPA and GATA2, have been recognised, but the culprit genes in the majority of AML pedigrees remain obscure. We applied a combined strategy of linkage analysis and next-generation sequencing (NGS) technology in an autosomal-dominant AML Chinese family with 11 cases in four generations. A genome-wide linkage scan using a 500K SNP genotyping array was conducted to identify a previously unreported candidate region on 20p13 with a maximum multipoint heterogeneity LOD (HLOD) score of 3.56 (P=0.00005). Targeted NGS within this region and whole-exome sequencing (WES) revealed a missense mutation in TGM6 (RefSeq, NM_198994.2:c.1550T>G, p.(L517W)), which cosegregated with the phenotype in this family, and was absent in 530 healthy controls. The mutated amino acid was located in a highly conserved position, which may be deleterious and affect the activation of TGM6. Our results strongly support the candidacy of TGM6 as a novel familial AML-associated gene.

Confirmation of papillary thyroid cancer susceptibility loci identified by genome-wide association studies of chromosomes 14q13, 9q22, 2q35 and 8p12 in a Chinese population.

BACKGROUND: Five single nucleotide polymorphisms (SNPs) were previously reported to be associated with thyroid cancer in European populations in two genome-wide association studies (GWAS): rs965513 (9q22.33), rs944289 (14q13.3), rs116909374 (14q13.3), rs966423 (2q35) and rs2439302 (8p12). Only the first two SNPs have been validated in independent populations and none were replicated in Chinese populations. METHODS: The above five SNPs were genotyped in 845 papillary thyroid cancer (PTC) and 503 benign thyroid tumour (BN) patients and 1005 controls in a Chinese population using the SNaPshot multiplex single nucleotide extension system. RESULTS: Significant associations were detected among PTC and rs944289 (p=8.007e-11), rs965513 (p=1.013e-4), rs966423 (p=1.688e-3) and rs2439302 (p=1.096e-4) in a dominant model, while the rs116909374 SNP was not detected in the Chinese population. The PTC risk increased with rise in accumulative numbers of risk alleles carried by individuals (p=5.929e-13). The PTC OR of carriers of six risk alleles (1.4% of the control population) was 23.587 compared with non-risk homozygotes (1.0% of the control population, with zero risk alleles). No individuals were homozygous for all the four SNPs (carriers of eight risk alleles) and only three PTC cases were carriers of seven risk alleles. A significant association between 14q13.3 SNP rs944289T and BN was also found (p=0.0014). CONCLUSIONS: Four candidate loci, rs965513 (9q22.33), rs944289 (14q13.3), rs966423 (2q35) and rs2439302 (8p12), identified by GWAS for PTC risk were confirmed in a Chinese population. The PTC risk of accumulative risk allele carriers increased with the number of risk alleles.

[Correlation of the methylation status of CpG islands in the promoter region of 10 genes with the 5-Fu chemosensitivity in 3 breast cancer cell lines].

OBJECTIVE: To explore the relationship between the methylation status of CpG islands in the promoter region of 10 genes in breast cancer cells and their sensitivity to 5-fluouracil (5-Fu), and to identify the genes responsible for the 5-Fu resistance in breast cancer. METHODS: Three cell lines (differently resistant to chemotherapy) were used in this study: Bcap-37 (IC(50): 289.77 microg/ml), T47D (IC(50): 134.16 microg/ml) and ZR-75-30 (IC(50): 4.20 microg/ml). The methylation profile of 10 genes (BAG1, C11ORF31, CBR1, CBR4, GJA1, FOXL2, IGFBP6, P4HA1, SRI and TYMS) in the 3 breast cancer cell lines was determined by methylation specific PCR. The steady-state mRNAs of ABCC8, CHFR and IGFBP6 genes were quantified by real-time RT PCR analysis. RESULTS: Among the 10 genes, only genes IGFBP6 and FOXL2 displayed differential DNA methylation pattern between the 5-Fu-resistant and 5-Fu-sensitive cell lines. The mRNA expression level of genes PRSS21, LOX, IGFBP6, ABCC8 and CHFR was quantified by real-time RT-PCR analysis. Except for CHFR, the expression level of the other 4 genes was correlated with the methylation status of CpG islands, namely, a lower expression level with methylation status and a higher level with demethylation status. CONCLUSION: The results of the present study have demonstrated that there are 8 genes with differential methylation status in chemosensitive and chemoresistant breast cancer cell lines, i.e. two genes more than the six genes we reported previously. Our findings provide both mechanistic insights for the drug resistance of breast cancer and the basis for further studies on potential application of the DNA methylation in this set of genes for prediction of chemosensitivity of breast cancer.