Research

This includes past or ongoing research projects as well as other interests I intend to pursue in the near future. These works can be broadly classified into application- and development-oriented projects.

This section will be updated periodically.

An Integrated Multi-Omics Analysis of Astrocyte–Immune Interactions in Glioblastoma

This study investigates astrocyte–immune cell interactions using an integrated multi-omics approach, combining transcriptomic, spatial transcriptomic/proteomic, and spatial analyses to define cellular crosstalk within the glioblastoma tumor microenvironment.

Characterizing MAIT Cells in Glioblastoma

We used TCR sequencing, single-cell RNA sequencing (scRNA-seq), and multiplex tissue imaging (CODEX) to investigate the role of mucosal-associated invariant T (MAIT) cells in glioblastoma (GBM). MAIT cells were present within GBM tumors and associated with immunosuppressive myeloid cells. Increased MAIT cell abundance correlated with poor patient survival in the TCGA-GBM cohort. Functional analyses suggest an immunosuppressive role mediated through IL-17 signaling.

Highlights

MAIT cell gene signatures correlate with poor survival in GBM
MAIT17 cells are the dominant MAIT subset in GBM tumors
MAIT cells contribute to immunosuppression via tumor-associated neutrophils

Publication:
Keretsu, S., Hana, T., Lee, A., Kedei, N., Malik, N., Kim, H., … Terabe, M. (2025). Characterization of MAIT cells in glioblastoma reveals a potential immunosuppressive role through a MAIT–Neutrophil axis. Neuro-Oncology Advances, 7(1), vdaf173.

Read full article

More details regarding my multi-omics projects will be updated soon.

Virtual Screening of MEROPS for SARS-CoV-2 Main Protease Inhibitors

Using the crystal structure of SARS-CoV-2 3CLpro (PDB: 6LU7), we identified potential inhibitors from the MEROPS protease agonist and antagonist database. Several candidates showed higher predicted binding affinity than the co-crystallized inhibitor.

Publication:
Keretsu, S., Bhujbal, S. P., & Cho, S. J. (2020). Rational approach toward COVID-19 main protease inhibitors via molecular docking, molecular dynamics simulation and free energy calculation. Scientific Reports, 10, 17716.

Read full article

Identification of Protein Complexes in Protein-Protein Interaction (PPI) Networks

Proteins rarely act alone in the cell. Instead, they work together in groups called protein complexes to carry out essential biological functions. Understanding how these complexes form and function is important for advancing basic biology and improving our knowledge of disease mechanisms. Scientists often use protein–protein interaction networks to identify protein complexes, but many existing methods assume these networks are static and focus mainly on finding tightly connected groups of proteins. As a result, some important complexes may be missed. In this study, we present a new computational approach called Weighted Edge-based Clustering (WEC) to improve the detection of protein complexes. Rather than relying only on how many connections proteins have, WEC evaluates the strength of interactions between proteins by combining information about network structure with patterns of gene expression. This allows the method to identify groups of proteins that are both strongly connected and biologically coordinated. When tested on multiple real-world biological datasets, WEC identified protein complexes more effectively than several widely used existing methods. By better reflecting the dynamic and functional nature of protein interactions, this approach provides a valuable tool for studying cellular organization and may support future research into the molecular basis of health and disease.

Algorithm 1

Algorithm for WEC

Input: PPI network: G(V,E), GEP: G(V);

Threshold Parameters: T_w, T_f, T_enrich and λ

Output: Set of Predicted Complex FC

Initialisation:

1:
FC = ∅
2:
PC = ∅
3:
for vertex v ∈ V do
4:
C = ∅;
5:
C ← v;
6:
for neighbor vertex u ∈ Nv do
7:
PCC = evaluate PCC(u, v);
8:
ECC = evaluate ECC(u, v);
9:
WEdge = (ECC × λ) + ((1 − λ) × PCC)
10:
if WEdge ≥ Tw then
11:
C ← u;
12:
end if
13:
end for
14:

        Cmax = arg max OS(C′, C)
         (C′ ∈ PC)
      
15:
if OS(Cmax, C) < Tf then
16:
PC ← C;
17:
else
18:

        if
        ( |∑(u,v)∈C ECC(u,v)|
        ≥ |∑(u′,v′)∈Cmax ECC(u′,v′)| )
        then
      
19:
PC ← C;
20:
discard C′;
21:
end if
22:
end if
23:
end for
24:
for C ∈ PC do
25:

        Uenrich =
        { u | u ∈ V, u ∉ C,
        e(u,v) ∈ E, v ∈ C };
      
26:
for u′ ∈ Uenrich do
27:
if NA(u′, C) ≥ Tenrich then
28:
C = C ∪ u′;
29:
end if
30:
end for
31:
FC ← C;
32:
end for

Publication:
Keretsu, S., & Sarmah, R. (2016). Weighted edge based clustering to identify protein complexes in protein–protein interaction networks incorporating gene expression profile. Computational Biology and Chemistry, 65, 69–79

Read full article

Computational Study of Serotonin Receptors (GPCR)

Recent studies demonstrate that Gaussian accelerated molecular dynamics (GaMD) can effectively probe binding and activation mechanisms of GPCRs. This work explores whether enhanced sampling methods can resolve unresolved structural questions in serotonin receptors.

Mechanism of STING Agonism

STING activation is critical for innate immune signaling and cancer immunotherapy. This project focuses on mechanistic understanding and rational design of hydrolysis-resistant STING agonists using computational drug discovery approaches. …More