Virtual Library

Start Your Search

Younghun Han



Author of

  • +

    MA10 - Emerging Technologies for Lung Cancer Detection (ID 129)

    • Event: WCLC 2019
    • Type: Mini Oral Session
    • Track: Screening and Early Detection
    • Presentations: 1
    • Now Available
    • +

      MA10.07 - Integrative Analysis of Epistasis Involving Oncogenesis-Related Genes in Lung Cancer Risk Development (Now Available) (ID 2502)

      15:15 - 16:45  |  Author(s): Younghun Han

      • Abstract
      • Presentation
      • Slides

      Background

      Our previous study identified significant genetic interactions within oncogenesis-related genes in lung cancer risk development. More genetic interactions may exist between oncogenesis-related genes and outside regions in the genome. A functional annotation and pathway analysis of the identified epistasis-related genes will advance our understanding about the complicated biological mechanisms underlying lung tumorigenesis.

      Method

      The genotypes from two independent lung cancer GWAS studies including a total of 23,351 lung cancer patients and 19,657 health controls with European ancestry were collected for the analysis. Pairwise epistasis was conducted between 27,722 SNPs, from 2,027 oncogenesis-related genes, and 317,624 SNPs from the rest of the genome. A two-stage study design including discovery and replication studies, and stringent Bonferroni correction for multiple statistical analysis were applied in the analysis, Additional genotyping and gene expression data from 409 independent individuals with Caucasian ancestry were used to evaluate the effect of identified epistasis on gene expression levels. The epistasis-involved genes, were submitted to DAVID, Reactome, and GeneMANIA for gene functional annotation and pathway analysis.

      Result

      Significant genetic interactions were identified between SNPs in gene pairs ATR-GALNT18 (Interaction OR=0.76, p value=7.98x10-13) and MET-DPF3 (Interaction OR=0.76, p value=1.62x10-12) in lung adenocarcinoma; and PICALM-PDZRN4 (Interaction OR=1.47, p value=1.67x10-12) in lung squamous carcinoma. None of these genes have been identified from previous main effect association studies in lung cancer. Further eQTL gene expression analysis revealed the significant association in expression levels between joint genotypes at rs637304:rs285581 and the PICALM gene expression (p=0.009). A total of 12 unique genes, from six significant interactions, including those from within oncogenesis-related genes and between oncogenesis-related genes and outside variants, were submitted to functional annotation and pathway analysis. Three of them (ATR, MET and FHIT) are shown to be related with lung cancer, and six of them (RAD51B, FHIT, CALNT18, RGL1, SYNE1 and TSPAN8) are involved in tobacco-use disorders. The top 10 pathways include TP53 regulates transcription of DNA repair genes (FDR=1.67x10-2), homologous DNA pairing and strand exchange (FDR=2.57x10-2), and Meiotic synapsis (3.08x10-2), etc. GeneMANIA predicted one gene network harboring all the 12 candidate genes, supporting the epistasis at 3 genes pairs and indirect interactions at 3 gene pairs.

      Conclusion

      We identified novel genes involved in lung cancer risk development by interacting with other genetic variants. The study provides evidence that epistasis explains part of the missing heritability in lung cancer; and complex gene network and pathways contribute to lung carcinogenesis.

      Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.

      Only Active Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login or select "Add to Cart" and proceed to checkout.

  • +

    P1.11 - Screening and Early Detection (ID 177)

    • Event: WCLC 2019
    • Type: Poster Viewing in the Exhibit Hall
    • Track: Screening and Early Detection
    • Presentations: 1
    • Moderators:
    • Coordinates: 9/08/2019, 09:45 - 18:00, Exhibit Hall
    • +

      P1.11-05 - Genetic Architecture of Lung Cancer Using Machine-Learning Approaches in Genome-Wide Association Studies (ID 547)

      09:45 - 18:00  |  Author(s): Younghun Han

      • Abstract
      • Slides

      Background

      Genome-wide association studies (GWAS) consisting of up to millions of single nucleotide polymorphisms (SNPs) have studied genetic influences to complex diseases and have identified thousands of associations. Few GWAS have explored interactions among SNPs that influence disease risks.

      Method

      Machine learning applications can define how SNPs jointly influence disease risks through interactions. Tree-based machine-learning applications; classification and regression trees (CART) and random forest (RF) methods are popular and convenient tools for understanding interactions influencing disease development. Here we apply these methods to elucidate the higher-order interactions that influence lung cancer risk. We applied tree-based approaches using 18,444 cases and 14,027 controls from lung cancer OncoArray GWAS data. We first selected the SNPs very significantly associated (p<0.00001) with lung cancer risk. RF, which consists of systematically fitting classification trees, was run 1,000 times to identify the most influential SNPs. Subsequently we applied CART to summarize and visualize interactions that predict risk.

      Result

      The final parsimonious tree included effects from genetic variants in CHRNA5, CLPTM1L, ZNRD1ASP, HCG9, TERT, CHRNB4, and DNAJC5 for overall lung. The final tree for adenocarcinoma lung showed the combination of genetic effects in or near ATM, CLPTLM1L, TERT, FSTL5, and DCTN4. The final tree for squamous cell carcinoma included CHRNA5, MRPL21, HLA, CASP8, and TAP2.

      Conclusion

      Our results confirmed associations with CHRAN5, TERT, and HLA observed in previous study (McKay et al., 2017). Machine learning methods in genomics provide some benefits over logistic regression model with respect to identifying subgroups at higher risk of lung cancer development on the basis of genetic characteristics.

      Only Active Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login or select "Add to Cart" and proceed to checkout.