Members

Ph.D. Students

EunyoungKim.jpg

Eunyoung Kim (김은영)

B.S. in Computer Science

eykim@gist.ac.kr | CV

Drug-Drug interaction and side effects prediction

Drug-Drug Interactions (DDIs) are important in drug discovery since it may lead to severe adverse drug reactions. Because of infeasibility of experiments on all drug combinations, I develop deep learning models for DDI prediction using drug features and transcriptomic data.

InGooLee.png

InGoo Lee (이인구)

B.S. in Interdisciplinary studies

dlsrnsladlek@gist.ac.kr | CV

Hit identification via deep learning

Identification of the hits that are possible drug candidates  is one of the important steps in drug discovery. However, chemical space is too vast to explore with chemical/biological experiments. Computational models for hit identification will reduce cost and accelerate drug discovery. Specifically, I hired a convolutional neural network (CNN) for a hit identification model, which captures important residue patterns in the protein sequence.

IljungJin.png

Iljung Jin (진일중)

B.S. in Computer Science and Information Engineering

skchin53@gist.ac.kr | CV

Cancer-Drug Response Prediction

Cancer patients show different responses on the treatment although they were treated with the same dose of the same drug; in some cases the therapy can aggravate the patient. So the identification of drug response on cancer is one of the important problems for precision medicine. I’m developing the deep learning model that can predict the cancer-drug response precisely and suggest MoA of cancer-drug response.

DohyunKim.jpg

Dohyun Kim (김도현)

B.S. in Computer Science

dhgold4u@gist.ac.kr | CV

Cancer Prognosis Prediction

Cancer prognosis is important to decide which treatment should be applied to the patient. If we are able to predict the accurate prognosis of cancer patients, we can recommend and offer appropriate therapies or drugs to patients. I'm interested in deep learning based cancer prognosis prediction model construction and robust prognostic feature selection.

HyunHoKim.jpg

Hyunho Kim (김현호)

B.S. in Computer Science

hyunhokim@gist.ac.kr | CV

Compound toxicity prediction and optimization

The research field that I have been interested in is compound toxicity prediction and optimization through deep learning methods. Since the numerous experimental databases are available, data-driven methods could be a cost-effective and less-biased solution. I’m currently focusing on developing hERG blocker prediction (Cardiotoxicity) using interpretable deep learning methods.

SongyeonLee.jpg

Songyeon Lee (이송연)

B.S. in Information Technology Engineering

songyeon@gist.ac.kr | CV

Cancer specific metabolic biomarker identification

For supporting rapid cancer growth and metastasis, some specific metabolic pathways are altered resulting in differentiation of the metabolites' concentration. To find metabolic biomarker candidates that represent the state of cancer and diagnosis, I'm applying the statistical methods and metabolic pathway analysis.

BongsungBae.jpg

Bongsung Bae (배봉성)

B.S. in Computer Science

bsbae402@gist.ac.kr

De Novo Optimization

I’m interested in the problem of de novo molecular optimization in computational drug design. In terms of methods, I am studying deep generative modeling and reinforcement learning for data-driven optimization.

Hansol.jpg

Hansol Lee (이한솔)

B.S. in Statistics

hansol.lee@gist.ac.kr

Antimicrobial peptide prediction

Bacteria are always evolving and becoming antibiotic resistant due to high rates of mutations in their DNA. Antimicrobial peptides (AMPs) kill invasive bacteria through non-specific mechanisms, and compared to conventional drugs, AMPs have shown a lower likelihood for bacteria to form resistance to. My goal is to facilitate the drug discovery process of AMP-influenced antibiotics through deep learning recognition of AMPs using protein data.

HaeleeBae.jpg

Haelee Bae (배해리)

B.E. in Computer Science and Engineering

haeleeeeleah@gist.ac.kr

Hit identification via deep learning

Identification of hit is important in drug discovery. Because there are many compounds and proteins, It takes much time and cost to find hit identification with experiment. Computation method can help to discover whether the compound and protein have interaction or not. I am interested in applying deep learning based data driven features to hit identification.

minsupark.jpg

Minsu Park (박민수)

B.S. in Chemistry

15pms@gist.ac.kr

Compound Membrane permeability prediction

I’m interested in predicting permeability of drugs through the biological membranes. Drugs need to pass biological membranes to reach their target, permeability prediction is important  for the drug efficacy. Membrane permeation is highly associated with molecular solvation property, I am studying about solvation property prediction with deep learning methods.