• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.12
• /
• pp.3383-3397
• /
• 2023

Scene graphs serve as semantic abstractions of images and play a crucial role in enhancing visual comprehension and reasoning. However, the performance of Scene Graph Generation is often compromised when working with biased data in real-world situations. While many existing systems focus on a single stage of learning for both feature extraction and classification, some employ Class-Balancing strategies, such as Re-weighting, Data Resampling, and Transfer Learning from head to tail. In this paper, we propose a novel approach that decouples the feature extraction and classification phases of the scene graph generation process. For feature extraction, we leverage a transformer-based architecture and design an adaptive calibration function specifically for predicate classification. This function enables us to dynamically adjust the classification scores for each predicate category. Additionally, we introduce a Distribution Alignment technique that effectively balances the class distribution after the feature extraction phase reaches a stable state, thereby facilitating the retraining of the classification head. Importantly, our Distribution Alignment strategy is model-independent and does not require additional supervision, making it applicable to a wide range of SGG models. Using the scene graph diagnostic toolkit on Visual Genome and several popular models, we achieved significant improvements over the previous state-of-the-art methods with our model. Compared to the TDE model, our model improved mR@100 by 70.5% for PredCls, by 84.0% for SGCls, and by 97.6% for SGDet tasks.

• Genomics & Informatics
• /
• v.21 no.4
• /
• pp.48.1-48.8
• /
• 2023

Liver cancer is the fourth leading cause of death worldwide. Well-known risk factors include hepatitis B virus and hepatitis C virus, along with exposure to aflatoxins, excessive alcohol consumption, obesity, and type 2 diabetes. Genomic variants play a crucial role in mediating the associations between these risk factors and liver cancer. However, the specific variants involved in this process remain under-explored. This study utilized a bioinformatics approach to identify genetic variants associated with liver cancer from various continents. Single-nucleotide polymorphisms associated with liver cancer were retrieved from the genome-wide association studies catalog. Prioritization was then performed using functional annotation with HaploReg v4.1 and the Ensembl database. The prevalence and allele frequencies of each variant were evaluated using Pearson correlation coefficients. Two variants, rs2294915 and rs2896019, encoded by the PNPLA3 gene, were found to be highly expressed in the liver tissue, as well as in the skin, cell-cultured fibroblasts, and adipose-subcutaneous tissue, all of which contribute to the risk of liver cancer. We further found that these two SNPs (rs2294915 and rs2896019) were positively correlated with the prevalence rate. Positive associations with the prevalence rate were more frequent in East Asian and African populations. We highlight the utility of this population-specific PNPLA3 genetic variant for genetic association studies and for the early prognosis and treatment of liver cancer. This study highlights the potential of integrating genomic databases with bioinformatic analysis to identify genetic variations involved in the pathogenesis of liver cancer. The genetic variants investigated in this study are likely to predispose to liver cancer and could affect its progression and aggressiveness. We recommend future research prioritizing the validation of these variations in clinical settings.

• Animal Bioscience
• /
• v.37 no.8
• /
• pp.1317-1332
• /
• 2024

Objective: Rare study of the non-coding and regulatory regions of the genome limits our ability to decode the mechanisms of fatty liver hemorrhage syndrome (FLHS) in chickens. Methods: Herein, we constructed the high-fat diet-induced FLHS chicken model to investigate the genome-wide active enhancers and transcriptome by H3K27ac target chromatin immunoprecipitation sequencing (ChIP-seq) and RNA sequencing (RNA-Seq) profiles of normal and FLHS liver tissues. Concurrently, an integrative analysis combining ChIP-seq with RNA-Seq and a comparative analysis with chicken FLHS, rat non-alcoholic fatty liver disease (NAFLD) and human NAFLD at the transcriptome level revealed the enhancer and super enhancer target genes and conservative genes involved in metabolic processes. Results: In total, 56 and 199 peak-genes were identified in upregulated peak-genes positively regulated by H3K27ac (Cor (peak-gene correlation) ≥0.5 and log2(FoldChange) ≥1) (PP) and downregulated peak-genes positively regulated by H3K27ac (Cor (peak-gene correlation) ≥0.5 and log2(FoldChange)≤-1) (PN), respectively; then we screened key regulatory targets mainly distributing in lipid metabolism (PCK1, APOA4, APOA1, INHBE) and apoptosis (KIT, NTRK2) together with MAPK and PPAR signaling pathway in FLHS. Intriguingly, PCK1 was also significantly covered in up-regulated super-enhancers (SEs), which further implied the vital role of PCK1 during the development of FLHS. Conclusion: Together, our studies have identified potential therapeutic biomarkers of PCK1 and elucidated novel insights into the pathogenesis of FLHS, especially for the epigenetic perspective.

• Journal of Life Science
• /
• v.25 no.3
• /
• pp.349-356
• /
• 2015

Thanks to recent advances in next-generation sequencing (NGS) technology, diverse livestock species have been dissected at the genome-wide sequence level. As for cattle, there are currently four Korean indigenous breeds registered with the Domestic Animal Diversity Information System of the Food and Agricultural Organization of the United Nations: Hanwoo, Chikso, Heugu, and Jeju Heugu. These native genetic resources were recently whole-genome resequenced using various NGS technologies, providing enormous single nucleotide polymorphism information across the genomes. The NGS application further provided biological such that Korean native cattle are genetically distant from some cattle breeds of European origins. In addition, the NGS technology was successfully applied to detect structural variations, particularly copy number variations that were usually difficult to identify at the genome-wide level with reasonable accuracy. Despite the success, those recent studies also showed an inherent limitation in sequencing only a representative individual of each breed. To elucidate the biological implications of the sequenced data, further confirmatory studies should be followed by sequencing or validating the population of each breed. Because NGS sequencing prices have consistently dropped, various population genomic theories can now be applied to the sequencing data obtained from the population of each breed of interest. There are still few such population studies available for the Korean native cattle breeds, but this situation will soon be improved with the recent initiative for NGS sequencing of diverse native livestock resources, including the Korean native cattle breeds.

• Journal of Microbiology and Biotechnology
• /
• v.18 no.2
• /
• pp.263-269
• /
• 2008

Hydrazinocurcumin (HC), a synthetic derivative of curcumin, has been reported to inhibit angiogenesis via unknown mechanisms. Understanding the molecular mechanisms of the drug's action is important for the development of improved compounds with better pharmacological properties. A genome-wide drug-induced haploinsufficiency screening of fission yeast gene deletion mutants has been applied to identify drug targets of HC. As a first step, the 50% inhibition concentration $(IC_{50})$ of HC was determined to be $2.2{\mu}M$. The initial screening of 4,158 mutants in 384-well plates using robotics was performed at concentrations of 2, 3, and $4{\mu}M$. A second screening was performed to detect sensitivity to HC on the plates. The first screening revealed 178 candidates, and the second screening resulted in 13 candidates, following the elimination of 165 false positives. Final filtering of the condition-dependent haploinsufficient genes gave eight target genes. Analysis of the specific targets of HC has shown that they are related to septum formation and the general transcription processes, which may be related to histone acetyltransferase. The target mutants showed 65% growth inhibition in response to HC compared with wild-type controls, as shown by liquid culture assay.

• Korean Journal of Microbiology
• /
• v.42 no.1
• /
• pp.1-6
• /
• 2006

Saccharomyces cerevisiae Dna2 helicase/endonuclease plays an essential role in removing DNA primers during Okazaki fragment processing in eukaryotic DNA replication. Genome-wide scale co-immunoprecipitation experiments predicted that Dna2 interacts with a novel protein YHR122W (1). In this study, we observed that overexpression of YHR122W gene suppressed the temperature-sensitive phenotype of $dna2\Delta405N$ mutation. To investigate direct interaction between these two proteins, a histidine-tagged recombinant YHR122W protein was expressed and purified from E. coli. Physical interaction between the purified YHR122W and Dna2 proteins was detected by enzyme-linked immunosorbent assays. Further more, the complex formation was most efficient at physiological salt concentration, 150 mM NaCl. The genetic and physical interactions between YHR122W and Dna2 shown in this study suggest that the biological functions of these two proteins may be closely related each other.

• Journal of Plant Biotechnology
• /
• v.45 no.1
• /
• pp.17-29
• /
• 2018

This study was conducted to investigate a morphological trait in 294 rice accessions including Korean breeding lines. We also carried out a genome-wide association study (GWAS) to detect significant single nucleotide polymorphism markers and candidate genes affecting major agronomic traits. A Manhattan plot analysis of GWAS using morphological traits showed that phenotypic and statistical significance was associated with a chromosome in each group. The significance of SNPs that were detected in this study was investigated by comparing them with those found previously studied QTL regions related to agronomic traits. As a result, SNP (S8-19815442), which is significant with regard to leaf angle, was located in the known QTL regions. To observe gene mutations related to leaf angle in a candidate gene, Os08g31950, its sequences were compared with sequences in previously selected rice varieties. In Os08g31950, a single nucleotide mutation occurred in one region. To compare relative RNA expression levels of candidate gene Os08g31950, obtained from GWAS analysis of 294 rice accessions and related to lateral leaf angle, we investigated relative levels by selecting 10 erect leaf angle varieties and 10 horizontal leaf angle varieties and examining real-time PCR. In Os08g31950, a high level of expression and various expression patterns were observed in all tissues. Also, Os08g31950 showed higher expression levels in the erect leaf angle variety group and higher expression rates in the leaf than in the root. The candidate gene detected through GWAS would be useful in developing new rice varieties with improved yield potential through future molecular breeding.

• Asian-Australasian Journal of Animal Sciences
• /
• v.33 no.10
• /
• pp.1558-1565
• /
• 2020

Objective: The objective of this study was to characterize the number of loci affecting growth traits and the distribution of single nucleotide polymorphism (SNP) effects on growth traits, and to understand the genetic architecture for growth traits in Hanwoo (Korean cattle) using genome-wide association study (GWAS), genomic partitioning, and hierarchical Bayesian mixture models. Methods: GWAS: A single-marker regression-based mixed model was used to test the association between SNPs and causal variants. A genotype relationship matrix was fitted as a random effect in this linear mixed model to correct the genetic structure of a sire family. Genomic restricted maximum likelihood and BayesR: A priori information included setting the fixed additive genetic variance to a pre-specified value; the first mixture component was set to zero, the second to 0.0001×σ²_g, the third 0.001×σ²_g, and the fourth to 0.01×σ²_g. BayesR fixed a priori information was not more than 1% of the genetic variance for each of the SNPs affecting the mixed distribution. Results: The GWAS revealed common genomic regions of 2 Mb on bovine chromosome 14 (BTA14) and 3 had a moderate effect that may contain causal variants for body weight at 6, 12, 18, and 24 months. This genomic region explained approximately 10% of the variance against total additive genetic variance and body weight heritability at 12, 18, and 24 months. BayesR identified the exact genomic region containing causal SNPs on BTA14, 3, and 22. However, the genetic variance explained by each chromosome or SNP was estimated to be very small compared to the total additive genetic variance. Causal SNPs for growth trait on BTA14 explained only 0.04% to 0.5% of the genetic variance Conclusion: Segregating mutations have a moderate effect on BTA14, 3, and 19; many other loci with small effects on growth traits at different ages were also identified.

• Journal of Plant Biotechnology
• /
• v.44 no.3
• /
• pp.264-270
• /
• 2017

Brassinosteroid (BR), a plant steroid hormone, plays a critical role in the growth and developmental processes through its canonical signaling and crosstalk with various internal and external signaling pathways. Recent studies have revealed the essential interplay mechanisms between BR and ABA during seed germination and early seedling establishment. However, molecular mechanisms for this important signaling crosstalk are largely unknown. To understand the crosstalk between BR-mediated signaling pathways and ABA functions during early seedling development, we carried out a comparative genome-wide transcriptome analysis with an Agilent Arabidopsis $4{\times}44K$ oligo chip. We selected and compared the expression patterns of ABA response genes in ABA-insensitive bes1-D mutant with wild type seedlings on which ABA was exogenously applied. As a result, we identified 2,353 significant differentially expressed genes (DEGs) in ABA-treated bes1-D and wild type seedlings. GO enrichment analysis revealed that ABA signaling, response, and metabolism were critically down-regulated by BR-activated signaling pathways. In addition, the genome-wide transcriptome analysis data revealed that BR-regulated signaling pathways were tightly connected to diverse signal cues including abiotic/biotic stress, auxin, ROS etc. In this study, we newly identified the molecular mechanisms of BR-mediated repression of ABA signaling outputs. Also, our data suggest that interplay among diverse signaling pathways is critical for the adaptive response of the plant to various environmental factors.

• Genes and Genomics
• /
• v.40 no.11
• /
• pp.1249-1258
• /
• 2018

The Jeju horse is an indigenous Korean horse breed that is currently registered with the Food and Agriculture Organization of the United Nations. However, there is severe lack of genomic studies on Jeju horse. This study was conducted to investigate genetic characteristics of horses including Jeju horse, Thoroughbred and Jeju crossbred (Jeju${\times}$Thoroughbred) populations. We compared the genomes of three horse populations using the Equine SNP70 Beadchip array. Short-range Linkage disequilibrium was the highest in Thoroughbred, whereas $r^2$ values were lowest in Jeju horse. Expected heterozygosity was the highest in Jeju crossbred (0.351), followed by the Thoroughbred (0.337) and Jeju horse (0.311). The level of inbreeding was slightly higher in Thoroughbred (-0.009) than in Jeju crossbred (-0.035) and Jeju horse (-0.038). $F_{ST}$ value was the highest between Jeju horse and Thoroughbred (0.113), whereas Jeju crossbred and Thoroughbred showed the lowest value (0.031). The genetic relationship was further assessed by principal component analysis, suggesting that Jeju crossbred is more genetically similar to Thoroughbred than Jeju horse population. Additionally, we detected potential selection signatures, for example, in loci located on LCORL/NCAPG and PROP1 genes that are known to influence body. Genome-wide analyses of the three horse populations showed that all the breeds had somewhat a low level of inbreeding within each population. In the population structure analysis, we found that Jeju crossbred was genetically closer to Thoroughbred than Jeju horse. Furthermore, we identified several signatures of selection which might be associated with traits of interest. To our current knowledge, this study is the first genomic research, analyzing genetic relationships of Jeju horse, Thoroughbred and Jeju crossbred.