Biology subjects

Hsieh, T.-T.

Publications and source records attributed to Hsieh, T.-T..

3 recordsLinked to original sources

bioRxiv2023

PGSbuilder: An end-to-end platform for human genome association analysis and polygenic risk score predictions

Understanding the genetic basis of human complex diseases is increasingly important in the development of precision medicine. Over the last decade, genome-wide association studies (GWAS) have become a key technique for detecting associations between common diseases and single nucleotide polymorphisms (SNPs) present in a cohort of individuals. Alternatively, the polygenic risk score (PRS), which often applies results from GWAS summary statistics, is calculated for the estimation of genetic propensity to a trait at the individual level. Despite many GWAS and PRS tools being available to analyze a large volume of genotype data, most clinicians and medical researchers are often not familiar with the bioinformatics tools and lack access to a high-performance computing cluster resource. To fill this gap, we provide a publicly available web server, PGSbuilder, for the GWAS and PRS analysis of human genomes with variant annotations. The user-friendly and intuitive PGSbuilder web server is developed to facilitate the discovery of the genetic variants associated with complex traits and diseases for medical professionals with limited computational skills. For GWAS analysis, PGSbuilder provides the most renowned analysis tool PLINK 2.0 package. For PRS, PGSbuilder provides six different PRS methods including Clumping and Thresholding, Lassosum, LDPred2, GenEpi, PRS-CS, and PRSice2. Furthermore, PGSbuilder provides an intuitive user interface to examine the annotated functional effects of variants from known biomedical databases and relevant literature using advanced natural language processing approaches. In conclusion, PGSbuilder offers a reliable platform to aid researchers in advancing the public perception of genomic risk and precision medicine for human disease genetics. PGSbuilder is freely accessible at http://pgsb.tw23.org.

bioinformatics↗

bioRxiv2020

accuEnhancer: Accurate enhancer prediction by integration of multiple cell type data with deep learning

Enhancers are one class of the regulatory elements that have been shown to act as key components to assist promoters in modulating the gene expression in living cells. At present, the number of enhancers as well as their activities in different cell types are still largely unclear. Previous studies have shown that enhancer activities are associated with various functional data, such as histone modifications, sequence motifs, and chromatin accessibilities. In this study, we utilized DNase data to build a deep learning model for predicting the H3K27ac peaks as the active enhancers in a target cell type. We propose joint training of multiple cell types to boost the model performance in predicting the enhancer activities of an unstudied cell type. The results demonstrated that by incorporating more datasets across different cell types, the complex regulatory patterns could be captured by deep learning models and the prediction accuracy can be largely improved. The analyses conducted in this study demonstrated that the cell type-specific enhancer activity can be predicted by joint learning of multiple cell type data using only DNase data and the primitive sequences as the input features. This reveals the importance of cross-cell type learning, and the constructed model can be applied to investigate potential active enhancers of a novel cell type which does not have the H3K27ac modification data yet. AvailabilityThe accuEnhancer package can be freely accessed at: https://github.com/callsobing/accuEnhancer

bioinformatics↗

bioRxiv2020

ezGeno: An Automatic Model Selection Package for Genomic Data Analysis

To facilitate the process of tailor-making a deep neural network for exploring the dynamics of genomic DNA, we have developed a hands-on package called ezGeno that automates the search process of various parameters and network structure. ezGeno considers three different sets of search spaces, namely, the number of filters, dilation factors, and the connectivity between different layers. ezGeno can be applied to any kind of 1D genomic input such as genomic sequences, histone modifications, DNase feature data and so on. Combinations of multiple abovementioned 1D features are also applicable. Specifically, for the task of predicting TF binding using genomic sequences as the input, ezGeno can consistently return the best performing set of parameters and network structure, as well as highlight the important segments within the original sequences. For the task of predicting tissue-specific enhancer activity using both sequence and DNase feature data as the input, ezGeno also regularly outperforms the hand-designed models. In this study, we demonstrate that ezGeno is superior in efficiency and accuracy when compared to AutoKeras, a general open-source AutoML package. The average AUC of ezGeno is also consistently higher than the result of using a one-layer DeepBind model. With the flexibility of ezGeno, we expect that this package can provide future researchers not only support of model design in their analysis of genomic studies but also more insights into the regulatory landscape. AvailabilityThe ezGeno package can be freely accessed at https://github.com/ailabstw/ezGeno. ContactDr. Chien-Yu Chen, chienyuchen@ntu.edu.tw

bioinformatics↗