Here is my CV; Contact: hanbin973@snu.ac.kr

News

I'm applying for a PhD in 2024. Feel free to contact me if we share any research interests.

Bio

I’m an undergraduate student at Department of Medicine, Seoul National University College of Medicine. My research interest primarily lies in the intersection of causal inference and population genetics. Currently, I work in Prof. Buhm Han laboratory. I’m largely interested in the philosophy of science and epistemology which makes me a strong proponent of the feminist standpoint theory.

In the mean time, I enjoy tennis and dining.

Education

Seoul National University

• Doctor of Medicine, Department of Medicine (2016 - current)
• Bachelor of Mathematics, Department of Mathematical Sciences (2016 - current)

---

Reserach interests

My current research goal is to promote identification as a central practice in scientific inquiry.

The role of casual inference

The biggest contribution of the modern study on causality was establishing the distinction between identification and estimation. In my opinion, recent biomedical research has focused largely on the latter while missing the essential components of identification. Instead of focusing on identification, literature tend to accept post-hoc interpretation of estimation outputs without proper justification. As a result, the meanings of research findings became very unclear and vague.

Why population genetics?

I believe that population genetics has a potential to overcome the current limitations in biomedical research. As one of the first field to adopt graph-based causal reasoning, causality has been a central component of the field. While most identification research focuses endogeneity issues such as establishing conditional exchangeability, my primary focus is on consistency and covariate distributions. For example, when questioning the causal role of body-mass index on a health outcome, one must first precisely define what really means to manipulate BMI which is a matter of consistency. Addressing such questions is possible only with the support of proper domain-specific knowledge and theory where population genetics is a very likely candidate.

Publications

• Journals

  • Hanbin Lee, and Buhm Han FastRNA: an efficient exact solution for PCA of single-cell RNA sequencing data based on a batch-accounting count model. American Journal of Human Genetics, in press, 2022
    Description: FastRNA produces batch-corrected low-dimensional representation of single-cell RNA sequencing data. We combine (1) the theory of sufficient statistics and (2) eigendecomposition to derive a completely sparse algorithm to perform feature selection and principal component analysis (PCA). The memory consumption and runtime are two to four orders of magnitude smaller than competing methods, enabling atlas-scale scRNA-seq data analysis within an ordinary personal computer.
    
    
  • Hanbin Lee, and Buhm Han A theory-based practical solution to correct for sex-differential participation bias. Genome Biology, 2022
    Description: The paper shows that sex-differential participation bias is largely dependent to the overall participation rate of a genetic cohort under standard parameteric assumptions. The major findings are (1) selection bias is highly sensitive to the parametric choice of the participation model (2) bias due to discrete variables contributing to study participation can be mitigated through stratification under low participation rate. As a result, it raises concern on the use of commercial genetic cohorts in which the target population is poorly specified, in turn, making the overall participation rate unidentifiable.
    
    
  • Chanwoo Kim*, Hanbin Lee*, Juhee Jeong, Keehoon Jung and Buhm Han MarcoPolo: a clustering-free approach to the exploration of differentially expressed genes along with group information in single-cell RNA-seq data. Nucleic Acids Research, 2022
    Description: MarcoPolo prioritizes potentially important genes in a single-cell RNA sequencing data through a unique ranking system. The method first fits a bimodal Poisson distribution and uses the fitted attributes to generate ranks of genes using four distinct metrics. The process does not require preceeding clustering of cells so it is robust to clustering results which are sensitive to program parameters.
    
    
  *: Equal contribution

• Unpublished Work

  • Moo Hyuk Lee*, Chanwoo Kim*, Buhm Han and Hanbin Lee Population genetic and counterfactual interpretation of genome-wide association study in admixed populations. In preparation, 2022
    Description: We propose a correct regression specification for association mapping in admixed populations through population genetics and causal inference. The theoretical properties of widely adopted estimators in admixed GWAS are discussed. We show that although the Armitage Trend Test (ATT) for quantitative traits is biased, the bias reduces as the strength of linkage disequilibrium (LD) increases and eventually vanishes under perfect linkage. This makes QATT the optimal test for discovering susceptibility loci. However, the bias can be severe when using genome-wide summary statistics produced by QATT to perform downstream analysis. We provide theoretically grounded guidelines for performing GWAS in admixed data and demonstrate it using the admixed samples in the UK Biobank.
    
    
  • Hanbin Lee, and Buhm Han Powerful and robust differnetially expressed gene analysis in single-cell RNA sequencing. In preparation, 2022
    Description: We demonsrate the power and robustness of Poisson regression in differentially expressed gene analysis (DEG) for single-cell RNA sequencing (scRNA-seq). We show that Poisson regression is (1) robust to any over or underdispersion of data, (2) effectively controls type 1 error rate, (3) powerful than pseudobulk methods and (4) fast and efficient enough to run on affordable computing systems.
    
    
  • Hanbin Lee, and Buhm Han A general fine-mapping framework for genome-wide association study through frequentist semiparametric regression. In preparation, 2022
    Description: The widely adopted Bayesian fine-mapping methods provide posterior distirbution of the likelihood for a given SNP to be the causal variant. Instead of providing the posterior distribution, we propose a framework that jointly estimates the position and the effect size of the causal variant using semiparametric regression. The framework can be naturally extended to include multiple ancestry and further improve the precision when admixed samples are availble.