Elephes Sung

Something of a scientist
Version v0.1.5
Updated
Author King of the Clingenland License MIT

About

Elephes Sung
print(“hello”)

I am currently a PhD student in Life Sciences at Imperial College, with a research interest in mathematical biophysics and immunology.

Previously, I completed an MRes in Systems and Synthetic Biology, also at Imperial College. Even prior to that, I got my bachelor’s degree in Nanomaterials and Nanotechnology.

Approximately 84% of people mispronounce my name on first reading, while 52.1% confidently assume it is Greek. Neither observation is especially concerning: you are entirely welcome to pronounce it however you wish.

You can email me at eu23@ic.ac.uk




Trajectory

This is a horizontal plot, scroll right to check it.

/*
            I need more quantitative                                                                                  
            & systematic methods to                                                                                   
            immunological mechanisms                                                                                  
                     │                                                                                                
  2019-2023          │       2023-2024                               2024-2028                                        
 ┌───────────────┐   │      ┌────────────────────┐                  ┌─────────────────────────────────────────┐       
 │ B.Eng.        ┼───┴─────►│ M.Res. Systems     ┼──────┬──────────►│ PhD in Life Sciences Research           │       
 │ Nanomaterials │          │ & Synthetic Biology│      │           │ President's Scholarship                 │       
 └┬──┬──────────┬┘          └─────────┬──────────┘      │           └──────┬──────────────────────────────────┘       
  │  └─────┬────┘                     │             in between:            │   Project 1. dynamics of interactions    
  │        │~3yrs                     │             ESA Hackthon           │          between killer immune cell      
  │        ▼                          ▼             Bayesian inference     │                     & target cells       
  │   UG project               Master's project       for astronauts'      │                                          
  │   Injectable Hydrogel      Agent-based modelling    health prediction  │   Project 2. dynamics of immune receptors
  │   NK cell delivery         NK cell-tumour                              │          & cell decision-making          
  │         cancer therapy             interactions                        │                                          
  ▼ ~1yr                                                                   │                                          
  lab intern                                                               small confidential project:                
  solar chemical                                                           Google Deepmind & Imperial College         
  battery                                                                                                             
┌──────────────────────────────────────────────────────────────────────────────────────────────────────────────┐      
│  TIMELINE                           TIMELINE                        TIMELINE                     TIMELINE    │      
└──────────────────────────────────────────────────────────────────────────────────────────────────────────────┘      
 */




Research and Development

My work connects experimental biology with quantitative modelling. These projects span Bayesian inference, research software, imaging analysis, simulation, and biomaterials development.


Bayesian inference for immunology

Research framework and open-source software

Graphic abstract showing how BARRACUDA converts time-lapse imaging into single-cell kill-contact histories and tests population heterogeneity and longitudinal variation.
BARRACUDA uses single-cell kill-contact histories to identify the mechanisms underlying variation in NK-cell cytotoxicity.

BARRACUDA

Bayesian Analysis Resolving Randomness and Alternative Causes Underlying Differential Activity is a quantitative framework for identifying why natural killer (NK) cell cytotoxic behaviour varies between individual cells. It combines event-count models with ordered contact-kill histories to distinguish stochastic encounters and killing decisions from stable cell-to-cell differences, donor effects, and changes induced by previous interactions.

Applied to time-lapse imaging of untreated, rituximab-treated, and CC-96673-treated NK cells, the framework found continuous heterogeneity across conditions while revealing distinct treatment mechanisms: rituximab primarily increased killing efficiency, whereas CC-96673 reduced cell-to-cell variability and produced a more homogeneous cytotoxic response.

Summary slide from Redshift, a hackathon project modelling astronaut immune health with cytokine dynamics and a normalized resilience factor.
Redshift: a compact cytokine-based immune-risk model for spaceflight.

Redshift

A brief Bayesian/ABC modelling project from the first European Space Agency hackathon, building a compact cytokine-based resilience factor for astronaut immune health.

Our team initially won, though the medal was later rescinded for administrative and political reasons.


Crappy softwares: imaging analysis and simulators

Single-cell dynamics from CytotoxicVision.

CytotoxicVision

A small but useful imaging pipeline for analysing time-lapse fluorescence microscopy data of NK-cell interactions with 721.221 target cells. Demo data from Dr Cathal Hosty from the Dan Davis lab.

The current version is still more of a demonstration framework than a polished package, but it already produces interpretable single-cell and population-level readouts.

Repository: github.com/ElephesSung/CytotoxicVision

Agent-based killer-target simulation from MantiShrimp.

MantiShrimp

A Python package for off-lattice, two-dimensional agent-based modelling of killer immune cell-target cell interactions. It combines cell migration and Hookean mechanics with contact formation, probabilistic killing, target death, and cell-state dynamics, and includes Bayesian event-count models for inference from per-cell contacts or kills.


Injectable Hydrogel

Up: Scanning electron microscopy (SEM) image showing the interconnected porous microstructure of the hydrogel scaffold.

← Injectable shape-memory hydrogel (gelatin) during compression and recovery.

↑ SEM view of the porous scaffold architecture.

porous cryogel

This was a biomaterials project I reproduced during my undergraduate studies, based on the injectable shape-memory scaffold reported by Bencherif and colleagues. The concept was simple but elegant: to create a porous hydrogel that could be compressed, injected through a syringe, and then recover its original structure after injection.

In their work, the porous architecture was generated through slow covalent crosslinking at low temperature. Ice crystals formed during the crosslinking process occupied space within the material and later served as pore templates.

Later, my colleagues and I developed a related approach for producing an ionically crosslinked porous alginate hydrogel. We first froze and lyophilised the alginate precursor, then immersed it in a calcium-ion solution to induce crosslinking. This produced a highly similar porous structure, while also preserving the material’s injectable behaviour.

Our original aim was to load NK cells into the hydrogel and use it as a delivery platform for solid tumours. Because of COVID and a few other unavoidable disruptions, I never managed to complete the bioengineering part of the project. It was a slightly frustrating ending.

Reference: Sidi A. Bencherif, R. Warren Sands, Deen Bhatta, et al. “Injectable preformed scaffolds with shape-memory properties.” Proceedings of the National Academy of Sciences 109(48): 19590-19595 (2012). doi.org/10.1073/pnas.1211516109




Publications

Google Scholar

Mathematical Biophysics

coming soon…

Nanomaterials & Drug Delivery

Review

Precision Medicine: The Road to In Vivo Synthetic Therapeutic Agent.

Yang‐Bao Miao, Zhao Wang, Fan-Xin Song, Renchi Gao, Zheng Deng, Guohui Zhang

Advanced Functional Materials, Vol. 35, Issue 47, p. 2510183 (2025)

Read more

Review

Recent Progress in Nanomaterial-Based Biosensors and Theranostic Nanomedicine for Bladder Cancer.

Fan-Xin Song, Xiaojian Xu, Hengze Ding, Le Yu, Haochen Huang, Jinting Hao, Chenghao Wu, Rui Liang, Shaohua Zhang

Biosensors, Vol. 13, Issue 1, p. 106 (2023)

Read more

Article

Achieving Precise Non-Invasive ROS Spatiotemporal Manipulation for Colon Cancer Immunotherapy.

Yang-Bao Miao, Hong-Xia Ren, Guohui Zhang, Fan-Xin Song, Weixin Liu, Yi Shi

Chemical Engineering Journal, Vol. 481, p. 148520 (2024)

Read more

Article

Tailoring a Luminescent Metal–Organic Framework Precise Inclusion of Pt-Aptamer Nanoparticle for Noninvasive Monitoring Parkinson’s Disease.

Yang-Bao Miao, Hong-Xia Ren, Qilong Zhong, Fan-Xin Song

Chemical Engineering Journal, Vol. 441, p. 136009 (2022)

Read more

Article

All-in-One, Solid-State, Solar-Powered Electrochemical Cell.

Yu Zhao, Chenyang Li, Fan-Xin Song, Yi Li, Yan Liu, Yajie Zhao, Xiaohong Zhang, Yu Zhao, Zhenhui Kang

ACS Applied Materials & Interfaces, Vol. 12, Issue 51, pp. 57182–57189 (2020)

Read more