Prospective Students

My lab focuses on exploiting specific molecular defects in cancer cells for targeted therapy. We are interested in translating the concepts of synthetic lethality to effective cancer treatments. In a broader sense, synthetic lethality describes functional interactions between two pathways/genes/mutations that synergistically reduce fitness and survival. We design CRISPR-based screens to identify synthetic lethal pathways and gene targets that can be manipulated to yield selective targeting of cancer cells. This work has expanded to include synthetic lethal interactions in DNA damage response pathways as part of our interests in deciphering the underlying connections between cancer initiation, progression and genomic instability. Recently, we have also identified new drug targets in metabolic pathways that paved the study for combination therapy with immune checkpoint inhibitors. In collaboration with clinicians and pathologists, our current work continues to espouse new cancer mechanisms in metabolic and DNA repair pathways for therapeutic applications.

Radiotherapy (RT) has an important role in the multimodality management of cancers and is employed in the neoadjuvant, definitive, or adjuvant setting. Mixed response to RT has been observed in the definitive setting and patients with poor response to RT have higher rate of recurrence and need for further salvage therapy, resulting in higher treatment morbidity. Further, distant metastasis and cancer relapse post-RT is also a major cause of treatment failure. Therefore, a pre-screening strategy to predict response to RT upfront would be valuable in identifying patients with radioresistant tumours as this would impact the treatment strategy at diagnosis. We recently identified a miRNA panel which predicts radioresistance and distant metastasis in cervical cancer through an unbiased microarray and miRNA screening approach. Analysis of public databases validated the prognostic value of the identified signature in determining outcome in cervical cancer. Given similarities in etiology and RT-based treatment strategies for cervical cancer and head and neck squamous cell carcinomas (HNSCC), we interrogated publicly available datasets for HNSCC and found that the identified miRNA signature also predicts poor outcome in HNSCC. The PhD candidate will participate in a multidiscipline research program with basic scientists and clinical scientists to (1) clinically validate the identified miRNA signature and their correlation to radioresistance and distant metastasis; (2) explore the potential of liquid biopsy-based non-invasive molecular surveillance and prediction strategies using identified molecular signatures in CC and HNSCC; (3) understand the mechanistic underpinnings of how candidate miRNAs can mediate the radioresistant and metastatic phenotypes. The work will advance efforts in the development of alternative treatment options to improve the prognosis of patients with resistant/recurrent cervical cancer and head and neck cancer.

The PhD candidate will participate in a multidiscipline research program with basic scientists, and clinical scientists to study the potential of targeting cysteine metabolism in ovarian clear cell carcinomas (OCCC), an aggressive and chemo-resistant tumor subtype comprising approximately 12-16% of epithelial ovarian cancers. Current treatment options for OCCC are limited to combination of platinum-based and other cytotoxic agents to which patients respond poorly due to intrinsic chemoresistance. We have previously reported that OCCC has profound dependence on cysteine for survival in vitro and in vivo, and cysteine-deprived glycolytic OCCC is abolished primarily by oxidative stress-dependent necrosis and ferroptosis. This suggests targeting cysteine metabolism as a novel alternative therapeutic strategy for OCCC. Our pilot study suggests depletion of the extracellular cysteine/cystine pool via pharmacological inhibition and engineered cysteine-degrading enzyme, can re-sensitise cisplatin-resistant cells to the cytotoxic effects of cisplatin.
We propose to study the effects of cysteine deprivation and its combinatory effects with cisplatin in 2D cell culture, 3D spheroid and patient-derived organoids and xenograft mouse models. We aim to use pharmacological inhibitors and cysteinase as a mode to systematically deplete extracellular cysteine for therapeutic applications as a cisplatin sensitizer, and to investigate the potential applicability of using novel delivery vehicle to deliver inhibitors or cysteinase to tumour sites for cysteine deprivation locally in in-vitro and in-vivo models. We will also elucidate molecular mechanism underlying the synergistic anti-cancer effect of cysteine deprivation and cisplatin in OCCC. The knowledge generated from this study will provide proof-of-concept data for the potential translational applications of targeted cysteine deprivation as a cisplatin re-sensitisation strategy for ovarian carcinoma which is the most lethal gynaecological cancer with majority of patients eventually becoming platinum-resistant with subsequent relapses.
The PhD candidate will participate in a multi-discipline research program with basic scientists, and clinical scientists to study the potential of targeting cysteine metabolism in ovarian clear cell carcinomas (OCCC), an aggressive and chemo-resistant tumor subtype comprising approximately 12-16% of epithelial ovarian cancers. Current treatment options for OCCC are limited to combination of platinum-based and other cytotoxic agents to which patients respond poorly due to intrinsic chemo-resistance. We have previously reported that OCCC has profound dependence on cysteine for survival in-vitro and in-vivo, and cysteine-deprived glycolytic OCCC is abolished primarily by oxidative stress-dependent necrosis and ferroptosis. This suggests targeting cysteine metabolism as a novel alternative therapeutic strategy for OCCC. Our pilot study suggests depletion of the extracellular cysteine/cystine pool via pharmacological inhibition and engineered cysteine-degrading enzyme, can re-sensitise cisplatin-resistant cells to the cytotoxic effects of cisplatin.
We propose to study the effects of cysteine deprivation and its combinatory effects with cisplatin in 2D cell culture, 3D spheroid and patient-derived organoids and xenograft mouse models. We aim to use pharmacological inhibitors and cysteinase as a mode to systematically deplete extracellular cysteine for therapeutic applications as a cisplatin sensitizer, and to investigate the potential applicability of using novel delivery vehicle to deliver inhibitors or cysteinase to tumour sites for cysteine deprivation locally in in-vitro and in-vivo models. We will also elucidate molecular mechanism underlying the synergistic anti-cancer effect of cysteine deprivation and cisplatin in OCCC. The knowledge generated from this study will provide proof-of-concept data for the potential translational applications of targeted cysteine deprivation as a cisplatin re-sensitisation strategy for ovarian carcinoma which is the most lethal gynaecological cancer with majority of patients eventually becoming platinum-resistant with subsequent relapses.

Ovarian cancer (OC) is the most lethal gynaecologic malignancy, with a 5-year survival rate of <30%. Immunotherapy is ineffective for OC due to an immunosuppressive microenvironment lacking tumor infiltrating leukocytes. Using a syngeneic mouse model, we observe that electrostimulation can inhibits tumor growth possibly through modulation of tumor microenvironment. The study aims to elucidate the underlying the molecular mechanism in a multidisciplinary team including cancer biologist, immunologist, and neurobiologist.

Emerging evidence implicates cancer/testis antigens (CTA) in cancer metabolism via their actions on the mitochondria. Our pilot study suggest CTA expression is involved in the metastatic potential of lung cancer. An increase in CTA in lung cancer brought about a more aggressive cancer phenotype, possibly through alterations in the cells’ mitochondria functions. This project aims to establish clinical association of CTA in lung cancer and study the underlying mechanism. The knowledge from this project is to better understand the role of CT antigens in cancer cell energetics and work towards using CTA in biomarker-driven clinical trials for mitochondria- or CTA-targeted therapeutics to provide more personalized treatment strategies to improve patient outcomes.

Current treatment for cervical cancer and head and neck cancer includes concurrent chemo-radiotherapy followed by brachytherapy. However, treatment failure characterized by recurrent/metastatic tumours is not uncommon. These tumours are often aggressive and resistant to current therapeutic strategies leaving these patients with little therapeutic options. In a pilot study, 43 first-biopsy samples from patients who were either disease-free (n=23) or recurrent/metastatic (n=20) after standard treatment, were subjected to an unbiased molecular screen and a recurrent/metastatic specific molecular signature was identified.
Several candidates in this signature have been suggested to have roles in maintaining the cancer stem cell (CSC) phenotype, which has been associated with disease recurrence/metastasis. However, their specific molecular functions remain to be elucidated and understanding how this signature affects the CSC population will enable the development of molecularly-targeted therapies for this subset of cells. Identification of our molecular signature from samples obtained at initial diagnosis suggests the exciting possibility of using this signature to predict possible relapse/metastatic disease at point of diagnosis. Together with molecular studies, the candidate is expected to apply this molecular signature to develop a liquid biopsy-based non-invasive molecular surveillance strategy. The long-term goal of this work aims to enable clinicians to predict recurrence/metastasis at initial diagnosis to better-tailor therapeutic regimes and prescribe them in a timelier manner to improve patient outcome. The project is part of a multidiscipline research program with basic scientists, translational researchers and clinical scientists.

Glycosaminoglycans are highly negatively charged molecules made up of repeating disaccharide subunits consisting of an amino sugar and an uronic acid. They are covalently linked to core protein backbones to form proteoglycans. Besides structural roles, glycosaminoglycans and proteoglycans have important biological functions in regulating cell behavior through their interactions with growth factors and behavior molecules. They are also involved in microRNA- and exosome-mediated regulation of cancer. In this study, we aim to elucidate the effects of these molecules on cancer cell activities, to investigate their potential clinical use as biomarkers and prognostic indicators, and to behavior them for the development of novel therapeutic targets. A variety of cell and molecular biology techniques will be employed in this project using both in-vitro and in-vivo models.

References:
1. Kumar KS et al (2021). FEBS J;288:486-506.
2. Tan XF et al (2019). Methods Mol Biol;1974:21-30.
3. Lucanus AJ et al (2018). Oncogene;37:833-838.
4. Iravani O et al (2017). Exp Cell Res;350:380-389.
5. Kumar AV et al (2014). Int J Cancer;135:2579-2592.
6. Ibrahim SA et al (2012). Int J Cancer;131:E884-896.
7. Yip GW (2011). Recent Pat Anticancer Drug Discov;6:164-165.

Positron Emission Tomography (PET) is a molecular imaging modality that allows visualization of molecular and metabolic-functional processes by the use of organic molecules and pharmaceuticals labeled with positron-emitting radionuclides (or tracers). These probes provide valuable insights into biochemical, physiological, pathological or pharmacological process in-vivo. PET tracers as molecular imaging probes have demonstrated significant clinical utility in oncology impacting on patient management in terms of early diagnosis, identifying histopathology, directing treatment, monitoring therapy response etc. Thus, molecular imaging tools will greatly contribute to the realization of modern precision oncology.
The PI, Prof Khong is the Director of The Clinical Imaging Research Centre (CIRC) of the Yong Loo Ling School of Medicine. CIRC is a core facility that for translational molecular imaging research that provides a platform for development and translation of molecular imaging biomarkers/tracers and theranostics to clinical practice. https://medicine.nus.edu.sg/circ/
Two important novel molecular imaging tracers for oncology will be developed at CIRC; 68Ga-Pentixafor (CXCR4 chemokine receptor) and 68Ga-FAPI (Fibroblast activation protein-α type II inhibitor). Both tracers are highly promising in the evaluation of various cancers using PET imaging, and early experience has found them to be advantageous to the currently used 18F-FDG (fluorodeoxyglucose) in multiple aspects with potential in filling unmet clinical needs.
We aim to evaluate its novel role of these tracers in the management of common cancers in Singapore.

To evaluate the clinical utility of a novel radiotracer, 68Ga-FAPI (Fibroblast activation protein-inhibitor) for PET imaging of Hepatocellular carcinoma (HCC); specifically, in its additional benefit to MRI alone, which is conventionally used as the imaging modality of choice for HCC, and as a combined hybrid imaging modality of PET/MRI.
We aim to study the added value of 68Ga-FAPI PET, combining it with MRI in a single hybrid modality, PET/MRI, in the context of the current diagnostic pathway in patients with or suspected to have HCC. Specifically, we aim to evaluate its role in (i) lesion detection and characterization prior to surgery (ii) systemic staging after radiological or histological confirmation of HCC (iii) characterization of indeterminate lesions on conventional MRI and (iv) work-up for the detection of HCC in patients with unexplained elevations of serum alpha-fetoprotein (AFP) concentration.

Personalized medicine has gained much momentum to become the mainstream practice to inform care to cancer patients. The basis of patient stratification relies on the existence of inter-tumoral heterogeneity. This has been extensively documented by The Cancer Genome Atlas (TCGA) projects identifying various molecular subtypes. These molecular subtypes have distinct clinicopathologic outcomes and hence are relevant for therapeutic intervention.
However, intra-tumoral heterogeneity (ITH) complicates this approach. Clonal diversity arising from ITH and the clonal evolution of alterations in the spatial and temporal context, results into multiple spatially separated subclones that would acquire distinct aberrations in the same gene, protein complex or signal transduction pathway. Moreover, the existence of heterogeneity is at the multi-omics levels. Therefore, leveraging on novel technologies to address the clonal diversity within ITH has become an urgent need. This project will dive into the spatial biology surrounding ITH to interrogate the tumor immune microenvironment (iTME) and the tumor epithelial-mesenchymal phenotypes (EMP). Cutting-edge technology platforms for high magnitude multiplexing signal detection with high resolution for precise spatial expression will be adopted. The multi-omics landscape from switches along the entire EMP spectrum would be associated with the remodeling of iTME in the selected cancer types and the precursors.

RNA based and RNA targeting behaviorize are increasingly becoming useful in clinical applications targeting various disease. Here we are trying to explore the relevance of one such class of RNA in relation to liver cancer. P65 and ADAR1 are also very important for liver homeostasis and have been associated with liver diseases including liver cancer. The proposal addresses an important link of how and what lncRNA are controlled by these 2 important proteins in cells which can be behavior later to target these molecules to treat liver cancer.
The student will be able to get a hands on training and skills related to conducting bench work in molecular and cell biology related expirements. By the end of the training he/she will be able to conceptualise behavior projects and design expirements and analyse the results obtained. Stronger students will also be trained to write papers and present data to others.

In this project, a high-throughput screening of small-molecule inhibitors targeting p52-ETS1 interaction will be conducted.
We will collaborate with industry partners to help us with the design and synthesis of the inhibitors, and also for further validation of our candidate inhibitors. Successful candidates can be patented and further developed for clinical trials.
The student will be able to get a hands-on training and skills related to conducting bench work in molecular and cell biology related experiments. By the end of the training he/she will be able to conceptualise research projects and design experiments and analyse the results obtained. Stronger students will also be trained to write papers and present data to others.

Messenger RNAs (mRNAs) comprise central protein-coding regions flanked by 5’ and 3’ untranslated regions (UTRs) that play important roles in post-transcriptional regulation. UTRs harbor many regulatory sequences and structures which control processes including mRNA stability, localization, export and translation efficiency. Intriguingly, the majority of mammalian genes generate multiple mRNA isoforms that differ in their UTRs. These alternative UTRs are produced by mechanisms including alternative promoter usage, splicing and alternative polyadenylation (APA). Critically, although independent efforts have begun to characterize each facet of UTR processing, little is known about how they converge to determine UTR heterogeneity. As UTRs are frequently subjected to multiple concurrent types of processing, a holistic approach is of key importance to understand how these processes collectively shape the cancer transcriptome and drive carcinogenesis. We anticipate that this work will lead to breakthroughs in our understanding of these diverse facets of RNA processing and their contribution to transcriptome heterogeneity and carcinogenesis, and identify novel regulators of critical cancer genes. Finally, we expect that 3’UTR variants and processing factors that are specifically dysregulated in cancer samples may represent promising new biomarkers for disease stratification, prediction of response to treatment, and/or lead to the development of novel targeted therapies.

Deep learning is a subset of machine learning which uses very deep neural networks to learn complex relationship between input variables and final output in very large datasets. In a process called supervised learning, an algorithm is trained iteratively on a high-quality annotated dataset. Each iteration culminates with a hypothetical/predicted output which is compared against the real output and the error is used intelligently to adjust the weight associated with each input parameter for the next iteration and the process is continued till the predicted output matches the real output. Finally, the trained model is used to predict the outcome for a new set of inputs. Working with bioinformaticians, clinicians and large healthcare data sets, you will create novel algorithms to identify patients at risk of drug adverse effects and who will benefit from precision treatments. You will be listed as an official member of the ethics approval and work with patient data to improve medical outcomes. A robust model of immediate translational value will be created to predict adverse reactions.

Redox chemistry is involved in literally every reaction that transfers an electron. The term oxidative-stress, refers to an imbalance in reductive potential for reactions that have a relevance to human biology and clinical outcomes. Despite the near universal acceptance that oxidative stress plays a fundamental role in aging, carcinogenesis and cardiovascular disease, a clinically actionable marker of oxidative stress remains to be discovered. We use mass spectrometry of known pathways, oxidatively modified substrates and precursor – product ratios to determine personalized signatures of oxidative stress in our collected patient cohorts. The goal of the project is to discover and characterize a biomarker that can be monitored in clinical use that guides the selection and dosage of anti-oxidant therapies. With success, the impact of this project can reshape the fields of nutrition and pharmaceutical science.

Technologies such as Blockchain, Digital Identification and Zero Knowledge Proofs are changing the possibilities for usage, study and sharing of medical information. This project is specific to those who want to make a difference in the lives of real patients thorough the application of digital mobility enablers in addition to decentralized (i.e. individually federated) analytics. In short, this project seeks to leverage on the previous work of the PI with World Health Organization, International Standards Organization (ISO) and major global healthcare networks to create a decentralized information sharing and analytics platform for real world data collection.

Can biochemical engineering fundamentally change the process of disease treatment and macromolecular creation? In this project you will use a novel form of nanoparticle invented by our lab (Nature Comms, PMID: 29129910) to study protein folding dynamics and novel cellular uptake mechanisms for drug delivery. The project is a continuation of research begun at Massachusetts Institute of Technology and will create nanometer scaled agents for the delivery of active protein substrates and for industrial manufacture of previously unproducable therapeutic solutions. Prof Drum was clinically trained at BWH, Harvard Med School and the team is composed of enthusiastic and highly talented scientists in training. Using this technology you will easily produce novel biological therapeutics and bioproduction sequences.
Expected outputs for the student will be at least one patent filing and 3-4 publications. You are invited to enquire if interested.

Bioorthogonal catalysis (BC) generates chemical reactions that are absent in normal physiological processes. With in situ drug synthesis, BC holds the promise of site-specific therapy with negligible systemic exposure or side effects. Previously, our lab reported the use of self-assembling, porous thermostable exoshells (tES) to encapsulate and deliver an iron-containing reaction centre for tumor regression in breast cancer mice. When administered with the prodrug indole-3-acetic acid, a natural plant product, the bioactive metabolites produced resulted in apoptosis of cancer cells and complete regression of tumors (Sadeghi et al. ACS Nano, 2022). Given the significance of oral delivery as a preferred mode for drug administration, we improved the stability of tES through multiple inter-subunit disulfide linkages. Preliminary studies have shown that ultra-stable tES protected the encapsulated BC against the adverse gastrointestinal environment and exhibited permeabilization through intestinal epithelial cells (Sadeghi et al. IJMS, 2022). Capitalizing on these recent findings, we aim to create a cost-effective oral prodrug therapy to treat an in vivo model of gastric cancer. The therapeutic strategy will couple the oral administration of the reactive BC centre followed by a benign natural substrate for in vivo evaluation together with safety profiling and biodistribution studies.

Nasopharyngeal carcinoma (NPC) is among top 10 cancer in males in Singapore. Locally advanced and metastatic disease have high relapse rate and poor prognosis even with best supportive care in combination with radiotherapy and chemotherapy. Attempts to identify novel targeted therapy for NPC are met with limited successes, with poor efficacy and high toxicity in clinical trials so far. From recent published data on genome wide sequencing of NPC patient samples, we identified that G1/S transition is among the top dysregulated pathways in NPC and the availability of FDA approved CDK4/6 inhibitors make this pathway an attractive molecular target in NPC. Through our preliminary data, global proteomic analysis revealed that Abemaciclib, one of CDK4/6 inhibitors, is able to modulate pathways in the innate immune system. In this project, the student will investigate the new Abemaciclib-based combination therapy to activate NK-cell mediated cytotoxicity against NPC. In addition, the precise molecular mechanism will be explored to identify novel treatment strategy in NPC.

My laboratory is behavior on transcription factors and gene regulation in cancer. A major effort by my laboratory focuses is listed here: (i) zinc finger protein 143 (ZNF143) is a key regulator in maintenance of progenitor cells integrity. ZNF143 secures essential chromosomal architectures with CTCF to maintain normal liver function, and alteration of such structures leads to hepatocellular carcinoma (HCC) formation. Here we propose to use an in vivo mouse model to demonstrate the importance of ZNF143 in preventing HCC formation. (ii) SALL4 is not expressed in most normal adult tissues and expressed in approximately 30% of cancers. Currently, there are no existing small molecule drugs targeting SALL4. In our laboratory we have discovered novel molecular glues that could degrade SALL4 in cancer cells and show in vivo efficacies. Further pre-clinical studies are needed to improve these chemical modalities. (iii) The human genome comprises of abundant regions that are not translated to proteins but actively transcribe long non-coding RNAs (ncRNAs). We aimed to discover novel lncRNAs that contribute to leukemogenesis in AML and are investigating the functional roles of these lncRNAs.

Approximately half of the human genome is constituted by repetitive sequences, which were once dubbed as ‘junk DNA’. These sequences play essential roles in the maintenance of genomic stability by regulating higher order chromatin organization, silencing of viral transposable, control of faithful chromosome segregation during cell division and heterochromatin formation. How the genomic stability of these repetitive elements is maintained is still unclear but the aberrant loss of transcriptional silencing at many repeats are correlated with carcinogenesis and aggressiveness in various cancers. Inappropriate transcription of these non-coding sequences is detrimental to cell health as it can induce DNA breakage and aberrant repair activation that subsequently result in deletion and fragmentation of the repeat DNA. Fragmented DNA eventually escapes into cytoplasm and blood to activate inflammatory 84 ehavior 84. A previous work revealed the presence of nuclear JAK2 kinase in several erythroleukemic cell lines, which aberrantly phosphorylate histone H3 tyrosine 41 (H3Y41p) to ectopically derepress oncogenes. Our work on the other hand detected H3Y41p on constitutive heterochromatin to control the transmission of epigenetic silencing of the underlying repeat sequences. We hypothesize that disruption of H3Y41p?dependent maintenance of heterochromatic integrity on repetitive elements can underlie dysregulation of DNA repeats that can induce inflammation and carcinogenesis. Hence identification of repeat maintenance dysregulation and fragmented repeat DNA can constitute possible early predictor of carcinogenesis. This project will study these mechanisms in leukemic oncogenesis. We will interrogate publicly curated acute myeloid leukemic (AML) patient genome databases to uncover mutations in possible epigenetic regulators critical for ensuring repeat stability. The consequences on repetitive DNA sequences and cell behavior due to a certain mutation will be subsequently analysed employing molecular and cellular analyses including animal models. We will establish methodologies to identify circulating DNA fragments in liquid biopsies from AML patients to correlate with mutations of the epigenetic factors. This work is expected to innovate generic technologies that will be applicable to interrogate genomic instability of repetitive DNA sequences as markers to predict oncogenic onset in other cancer models.

Ovarian cancers account for a significant proportion of gynaecological cancers worldwide, with a majority of patients presenting at advanced stages. Although initial response to treatment may be promising, relapse rates are high and 5-year survival is often poor.
Migration and invasion of cancer cells constitute fundamental processes in tumour progression and metastasis and are governed by alterations of various families of cytoskeletal proteins, such as microtubules, actin, and intermediate filaments. These cytoskeletal proteins act in a coordinated manner and mediate numerous cellular and biochemical processes that influence cell shape, structure, division, motility as well as interactions of cells with their extracellular environment. Emerging evidence indicates that cytoskeletal alterations in cancer cells derived from aggressive tumour subtypes render them resistant to killing by cytotoxic immune cells found in the tumour microenvironment.
Recent studies have shown that actin-rich structures are associated with increased migration and facilitates epithelial-to-mesenchymal transition in response to increased stiffness of the extracellular tumour microenvironment. However, mechanisms by which cancer cell cytoskeletal rearrangements and its roles in immune evasion are not fully understood.
The aim of this project is to examine the cytoskeletal mechanisms influencing ovarian tumour cell behavior, such as alterations in cell shape, cytoskeletal architecture and how these affect the interactions with the cells in the tumour microenvironment, such as immune cells. Analysis will be carried out by assessing the composition of actin fibres, cell-extracellular matrix adhesions, actin-rich migratory structures as well as cytotoxic properties of the relevant immune cells using immunofluorescent imaging, time-lapse microscopy, FACS, modulation of genetic expression of key cytoskeletal proteins, including gelsolin and other actin-binding proteins.

Lung cancer is a leading cause of cancer-related death globally. Metastasis is responsible for as much as 90% of cancer-associated mortality, yet it remains the most poorly understood component of cancer pathogenesis. The precise mechanism of tumor metastasis is still largely unknown. This field of research has garnered renewed interest ever since the discovery of vesicle-based cell-cell communication mediated by exosomes, particularly tumor-derived exosomes. Although multiple factors are known to contribute to cancer metastasis, epithelial-mesenchymal transition (EMT) has been implicated as the critical event initiating cancer invasion, intravasation and extravasation of tumor cells of the metastatic cascade.
Therefore, it is crucial to attempt the investigation of EMT driver genes to discover new signaling pathways and druggable targets for treatment of lung cancer, one of the top issues impacting the global health care sector. Based on our preliminary results, FAM3C (ILEI) has been found as a critical driver in the metastatic cascade of advanced lung cancer.
Mechanistic experiments further revealed that tumor-derived exosomes enriched FAM3C can activate EMT through Ras pathway effector RalA signaling in NSCLC cells to initiate tumour cell invasiveness and motility. Importantly, we established that RalA, a small GTPase, is a partner protein for FAM3C. These findings imply that targeting FAM3C-RalA interaction may be a feasible strategy to suppress cancer metastasis. In this project, the student will investigate association between FAM3C and Ras pathway effector RalA in driving tumor growth and metastasis of NSCLC. Next, a large-scale drug library screening will be conducted to identify novel and potent inhibitors in treatment of NSCLC through inhibiting FAM3C-RalA interaction. Furthermore, downstream signaling pathway will be investigated to understand FAM3C driven oncogenic mechanisms. Finally, a suitable xenograft mouse model will be used to validate the antimetastatic activity of the shortlisted inhibitors.

Chromatin-associated proteins are emerging as new drug targets for a variety of diseases, especially cancer. Our laboratory is interested to study how chromatin deficiencies underlie human diseases, with a strong focus on therapeutics development. We employ varied experimental approaches such as Next-Generation Sequencing, genome-editing, genetic/chemical screens as well as cancer models (PDXs, tumor organoids, etc.) to address our questions of interest. We recently conducted a systematic investigation of all chromatin regulators in cancer where we overlaid gene expression patterns with clinical data (metastasis, treatment success, and survival) followed by functional studies. This effort enabled us to prioritize candidate factors for downstream therapeutics development. The student will have opportunities to acquire strong expertise in various (epi)genomics and proteomics techniques, perform cancer assays and high-throughput screens, as well as acquire drug development knowledge.
The student will also work closely with our clinical collaborators, as well as drug development colleagues at A*STAR EDDC. For more information about the lab, please see: https://medicine.nus.edu.sg/phys/research/research-programs/cancer-programme/tee-wee-wei/

Drug resistance, either intrinsic or acquired, represents a major bugbear in precision medicine. It is increasing clear that epigenetic reprogramming mechanisms contribute to transcriptional plasticity in cancer cells enabling lineage transformation towards a ‘drug-tolerant persister’ (DTP) state. These rare population of DTP cancer cells acquire transcriptomic and epigenetic features resembling that of early embryonic cells that are developmentally plastic.
Accordingly, they survive and adapt readily to therapeutic pressures, and over time can re-populate the tumour. Using established commercial cell line and patient-derived xenograft and organoid models of resistance (e.g. Paclitaxel in breast and 2nd/3rd generation TKIs in lung cancers), we propose to isolate and characterize the DTP cancer cells to identify the epigenetic drivers of therapeutic resistance. We intend to also use CRISPR-based genetic screens and pharmacologic approaches to identify epigenetic regulators that impede the formation or progression of DTPs, and assess for improved clinical outcomes in combination with standard-of-care treatments. To this end, we have developed epigenetic technologies that are amenable for low input material and our preliminary efforts have identified potential candidates for downstream studies.
For more information about the lab, please see: https://medicine.nus.edu.sg/phys/research/research-programs/cancer-programme/tee-wee-wei/

Our lab has a long-standing interest in understanding the epigenetic mechanisms of lineage plasticity in early development (Hu et al., Nature Cell Biology 2020). More recently, we have broadened our investigations to study how cancer cells undergo epigenetic reprogramming to induce phenotypic plasticity. This cross-fertilization strategy has been rewarding, directing us to uncover conserved players/mechanisms implicated in epigenetic reprogramming and exposing novel cancer vulnerabilities.
Indeed cancer cells are known to exhibit certain embryonic features and may co-opt early developmental programs to drive phenotypic plasticity giving rise to disease progression and therapeutic resistance. In ongoing studies, we found that reactivation of select early-embryonic factors can indeed promote cancer cell phenotypic plasticity through epigenetic reprogramming mechanisms, similar to that occurring during early embryogenesis. In this project, the student will combine single-cell RNA sequencing (scRNA-seq), spatial profiling and genetic manipulations to systematically investigate the roles of early-embryonic factors in tumorigenesis and dissect how their heterogenous expressions may correlate to distinct cancer cell states through epigenetic mechanisms.
For more information about the lab, please see: https://medicine.nus.edu.sg/phys/research/research-programs/cancer-programme/tee-wee-wei/

Natural proteins have evolved over billions of years, yet evolution has explored only a small fraction of the enormous sequence, structure, and functional space available to proteins. Computational protein design seeks to go beyond nature by engineering novel protein sequences with new structures and functions. By enabling the creation of proteins and peptides with desired properties and activities, this approach holds great promise for developing next-generation therapeutics to modulate cellular pathways and treat major human diseases such as cancer and Alzheimer’s disease.

Recent advances in artificial intelligence have opened new opportunities for designing biomolecules and therapeutic compounds with desired structural and functional properties. Despite remarkable progress, major challenges remain in controlling molecular structure and interactions, achieving high affinity and specificity, optimizing multiple biochemical and biophysical properties simultaneously, and translating computational designs into experimentally validated therapeutic candidates.

In this project, we will develop next-generation computational methods for protein and peptide design and AI-driven drug discovery, exploring the integration of artificial intelligence with evolutionary information, physical modeling, molecular simulations, and experimental data.

The developed methods will be systematically evaluated using computational benchmarks and, where appropriate, experimentally validated through collaborations with biological, biomedical, and drug-discovery laboratories.

Projects can be tailored according to the student’s background, research interests, and duration of study.

Proteins and RNAs are among the most essential biomolecules in life. Their biological functions are closely linked to their three-dimensional structures, making accurate structure determination critical for understanding molecular mechanisms and biomedical applications. Although hundreds of millions of protein and RNA sequences have been identified, only a small fraction have experimentally determined three-dimensional structures. Computational structure prediction is therefore essential for bridging this sequence–structure gap.

Over the past decade, the field of biomolecular structure prediction has been transformed by artificial intelligence and deep learning, with major contributions from the PI and other groups. Despite this progress, important challenges remain. These include accurate modeling of multi-domain proteins and large biomolecular complexes, RNA structure prediction, conformational flexibility and dynamics, and the effective integration of evolutionary, physical, and experimental information.

In this project, we will develop next-generation computational methods for protein and RNA structure prediction, exploring the integration of artificial intelligence with evolutionary information, physical modeling, molecular simulations, and experimental structural data.

The developed methods will be systematically benchmarked against state-of-the-art approaches using established structural datasets and international blind experiments such as CASP, which provides a rigorous assessment of protein and RNA structure prediction methods.

Projects can be tailored according to the student’s background, research interests, and duration of study.

Phase separation-mediated DNA/RNA/protein condensates represents one of the most exciting and important findings in molecular and cell biology fields these years. These biomolecular condensates play crucial roles in transcription, RNA processing, translation, and signalling transduction. Their highly dynamic and heterogeneous properties in response to environmental perturbations and stress underscore their physiological significance.
We have developed and applied novel FANCI and FLASH methods to purify RNA condensates and study RNA-protein interactions. We plan to use these cutting-edge technologies to study RNA condensates in different cancer cells as well as in cancer cells subjected to different stress challenges such as chemotherapy and starvation. Next we will confirm the unique enrichment of candidate RNA and proteins in condensates from clinical samples. Genetic and biochemical tools will be applied to manipulate the abundance and localisation of promising proteins and RNAs within these condensates—a process I refer to as compartmentalised biochemistry—to explore their function in tumour survival and proliferation.
Detailed characterisation of the RNA and protein composition as well as RNA-protein interactions within RNA condensates will not only reveal the principles of RNA dynamics in biomolecular condensates but also enhance our understanding of tumour survival strategies and identify novel therapeutic targets.

High-grade serous ovarian cancer (HGSOC) is the deadliest gynecologic cancer, largely due to late diagnosis, rapid dissemination, and treatment resistance. A critical but understudied factor is how cancer cells adapt to the harsh mechanical forces within the peritoneal cavity to survive, spread, and resist therapy. This PhD project will investigate PTK6 as a master regulator of mechanoadaptation, helping tumor cells migrate through confined spaces, repair DNA damage, and withstand therapeutic stress. You will employ cutting-edge approaches including CRISPR gene editing, microfluidic confinement platforms, 3D extracellular matrix models, and live-cell imaging to dissect PTK6-driven signaling in real time. The project will explore how PTK6 controls cytoskeletal remodeling, focal adhesion dynamics, YAP-mediated mechanotransduction, and DNA repair following nuclear stress. Findings will be validated using patient-derived organoids to ensure translational relevance. This is an opportunity to work at the interface of mechanobiology, cancer signaling, and precision therapeutics in collaboration with leading mechanobiology researchers. The goal is to identify actionable vulnerabilities that can halt metastasis and overcome drug resistance in ovarian cancer. You will graduate with expertise in advanced experimental systems, translational cancer biology, and therapeutic target discovery.

Despite advances in chemotherapy, PARP inhibitors, and immunotherapy, most ovarian cancer patients relapse due to drug resistance and an immunosuppressive tumor microenvironment. This project explores a bold concept: targeting PTK6 to simultaneously enhance therapy-induced DNA damage and activate anti-tumor immunity. You will investigate how PTK6 inhibition disrupts DNA repair, leading to cytosolic DNA accumulation and activation of the cGAS–STING pathway, an innate immune signaling axis capable of transforming “cold” tumors into immune-responsive ones. The work will integrate molecular biology, tumor immunology, and translational therapeutics. Using CRISPR models, platinum and PARP inhibitor treatments, and orthotopic immunocompetent mouse systems, you will evaluate tumor response, immune cell infiltration, interferon signaling, and therapeutic synergy with checkpoint inhibitors. Multi-omics and immunoprofiling approaches will identify biomarkers predicting response to PTK6-targeted therapy. This project is ideal for students interested in next-generation cancer therapy at the intersection of DNA damage, immune activation, and precision medicine. The outcomes may redefine how targeted therapy and immunotherapy are combined in ovarian cancer.

Cervical cancer remains a major global health burden, with treatment resistance and metastasis limiting patient survival. Emerging evidence suggests that Lyn kinase, a Src family signaling node, may drive tumor progression, immune evasion, and therapy failure, yet its role in cervical cancer is largely unexplored. This PhD project will uncover how Lyn rewires oncogenic signaling and evaluate a novel Lyn-specific inhibitor as a precision therapeutic strategy. You will analyze Lyn expression in clinical datasets and models, dissect its role in proliferation, EMT, invasion, and immune interactions, and map downstream signaling networks using transcriptomics and phosphoproteomics. A major translational component will test a first-in-class Lyn kinase inhibitor in combination with chemotherapy to determine its ability to suppress tumor growth and restore treatment sensitivity. Preclinical validation in in vivo models will provide a pathway toward clinical translation. This project offers an opportunity to contribute to the development of a targeted therapy from discovery to preclinical validation. Graduates will gain deep expertise in cancer signaling, translational drug development, and precision oncology

Alternative Lengthening of Telomeres (ALT) is a recombination-based telomere maintenance mechanism used by a subset of cancers, including many sarcomas and subsets of brain tumours. Our laboratory is interested in understanding how a normal cell acquires the molecular competence to activate ALT during malignant transformation, which cellular pathways subsequently sustain this state, and whether these features can be exploited for cancer detection or therapy.
Projects may investigate how changes in telomeric chromatin, DNA repair, transcriptional regulation and nuclear organisation contribute to the emergence and maintenance of ALT. We are particularly interested in identifying molecular events that enable ALT activation and defining vulnerabilities that distinguish ALT-positive cancer cells from normal or telomerase-positive cells. Complementary translational directions include developing biomarkers and therapeutic strategies for ALT-positive cancers using patient-derived samples and proteomic approaches.
Depending on the student’s interests and background, projects may combine molecular and cell biology, genome engineering, microscopy, functional genomics, proteomics and patient-derived models. Rather than prescribing a single fixed project, we aim to develop a focused PhD project together with the student based on their scientific interests, emerging findings in the laboratory and the most promising questions at the time.

Neovascular age-related macular degeneration (nAMD) is a leading cause of irreversible vision loss in older adults, driven by choroidal neovascularization, inflammation and subretinal fibrosis. Although anti-VEGF therapies such as aflibercept have transformed care, many patients develop incomplete or diminishing responses, with approximately 45% developing subretinal fibrosis within two years despite continuous treatment. This highlights an urgent need to target VEGF-independent mechanisms driving disease progression. We have developed CHL-4, a first-in-class selective Lyn kinase inhibitor with potent anti-angiogenic and anti-fibrotic activity. We hypothesize that Lyn intersects with YAP/TEAD signaling, a mechanotransduction pathway linking pathological angiogenesis, inflammation, extracellular-matrix remodeling and fibrosis. Progressive retinal stiffening may establish a self-sustaining YAP/TEAD–fibrosis circuit that persists despite VEGF suppression. This study will determine whether CHL-4 suppresses YAP/TAZ activation and TEAD-dependent transcription and whether combined Lyn–YAP/TEAD inhibition provides superior control of angiogenesis, inflammation and EMT-mediated fibrosis. Mechanistic studies will evaluate YAP/TAZ nuclear localization, TEAD activity and downstream effectors including CTGF/CCN2, CYR61/CCN1, ANKRD1 and AXL. By simultaneously targeting vascular, inflammatory and fibrotic drivers, this study could position CHL-4, alone or combined with YAP/TEAD or anti-VEGF therapy, as a next-generation treatment for anti-VEGF-resistant nAMD.

Triple-negative breast cancer (TNBC) remains an aggressive disease with high rates of metastatic recurrence and treatment resistance, underscoring the urgent need for biomarker-guided therapies that eliminate tumor-driving vulnerabilities. We have identified DP103/DDX20 as a clinically relevant driver of TNBC progression, cancer stemness, metastasis and therapeutic resistance. However, its intracellular and multifunctional nature makes DP103 difficult to target using conventional therapeutic antibodies or small-molecule inhibitors.
This PhD project will develop a first-in-class tumor-targeted degrader–antibody conjugate (DP103-DAC) that combines the precision of antibody-mediated tumor recognition with PROTAC-mediated degradation of intracellular DP103. A DP103-selective PROTAC will first be developed and optimized for potent, sustained and selective target degradation. The degrader will subsequently be conjugated to an internalizing TNBC-targeting antibody, prioritizing TROP2, to selectively deliver the therapeutic payload into tumor cells. Lead candidates will be evaluated for suppression of cancer stemness, metastatic potential and treatment resistance using TNBC cell lines, patient-derived organoids and in vivo tumor models. By integrating tumor-surface antigen expression with DP103 dependency, this project aims to establish a dual-biomarker precision-treatment strategy with an improved therapeutic index and provide preclinical evidence for translating targeted DP103 degradation into a new treatment for patients with aggressive TNBC.

Emerging contaminants, including PFAS, organophosphate esters, plastic additives, pharmaceuticals, and environmental transformation products represent a growing but poorly defined threat to human health. The rapid expansion of chemical exposures has outpaced conventional toxicological testing, leaving many contaminants insufficiently characterized and their disease-causing mechanisms unknown. This PhD project will establish an AI-guided precision toxicology platform that transforms how emerging environmental contaminants are prioritized, mechanistically investigated, and evaluated for human-health risk. Interpretable machine-learning models will integrate chemical structure, environmental occurrence, human exposure, toxicity, molecular targets, biological pathways, and disease associations to identify high-priority contaminants and generate experimentally testable mechanistic hypotheses. Predictions will be validated using disease- and tissue-relevant human cell systems, followed by targeted molecular perturbation to establish causal pathways linking chemical exposure to toxicological and disease-relevant phenotypes. Only the strongest hypotheses will progress to higher-tier human organoid, three-dimensional culture, Drosophila, or focused in vivo models, providing a rigorous and ethically responsible evidence framework. By integrating AI prediction, human-relevant experimentation, mechanistic pharmacology, and cross-model validation, this research aims to move environmental toxicology from descriptive hazard identification toward mechanism-informed prediction of human disease risk, enabling earlier identification of hazardous chemicals and more effective prioritization for public-health intervention and regulatory assessment.

How do biomolecules recognize the correct partners among the enormous number of molecules in a cell? Molecular recognition underlies protein interactions, immune recognition, signaling, and therapeutic targeting, yet accurately predicting interaction geometry, affinity, and specificity remains a major challenge. Recent advances in protein and RNA language models and structural AI provide new opportunities to move beyond pairwise interaction prediction toward a general framework for molecular recognition.

In this project, we will develop new AI methods that integrate sequence, three-dimensional structure, evolutionary information, physical interactions, and biological context to predict and understand biomolecular recognition. Applications may include protein–protein, protein–RNA, antibody–antigen, and TCR–pMHC interactions. A central goal is to connect atomic-level interaction interfaces and binding specificity with large-scale interaction networks, enabling mechanistic understanding of molecular recognition and applications in immunology and therapeutic discovery.

The developed methods will be systematically evaluated on benchmark datasets for interaction geometry, affinity, and specificity and, where appropriate, experimentally validated through collaborations with biological and biomedical laboratories.

Projects can be tailored according to the student's background, research interests, and duration of study.

Knowing the molecular function of a protein is not sufficient to explain its biological role. The same protein may participate in different interactions, pathways, and cellular processes depending on cell type, tissue environment, disease state, and molecular context. Most existing protein-function models, however, predict static annotations from sequence or structure and largely ignore this context dependence.

In this project, we will develop new AI methods that move beyond conventional function prediction toward context-aware modeling of protein function and biological mechanisms. By integrating protein sequence and structure with molecular interactions, biomedical knowledge, transcriptomic and spatial data, the models will seek to connect molecular properties to pathways, cellular states, tissues, and disease phenotypes. Applications may include protein and enzyme function, pathway and mechanism discovery, spatial transcriptomics, and tumor microenvironment analysis. The long-term goal is to explain not only what a protein can do, but when, where, and through which mechanism it performs a biological role.

The developed methods will be systematically evaluated using established functional benchmarks and independent biological datasets, with particular emphasis on their ability to generalize across proteins, cellular contexts, and disease conditions.

Projects can be tailored according to the student’s background, research interests, and duration of study.

Biological information is distributed across multiple interconnected levels: sequence determines structural possibilities, structure shapes molecular interactions, interactions contribute to function, and function depends on cellular and disease context. Yet most current AI models learn only one molecular type, data modality, or prediction task, leaving these levels fragmented across separate representations and limiting our ability to reason across biological scales.

This project asks whether a unified AI model can learn the principles connecting sequence, structure, interaction, function, and biological context across proteins, RNAs, and small molecules. We will develop multimodal foundation models that integrate molecular sequences, three-dimensional structures, physical interactions, biological annotations, and other relevant data within shared representations. Rather than building separate models for individual tasks, the goal is to develop transferable models capable of reasoning across molecular types and scales. Such models could provide a common foundation for structure prediction, molecular interaction modeling, functional understanding, and molecular design.

The developed models will be systematically evaluated across diverse downstream tasks, with particular emphasis on cross-modal transfer, generalization to unseen molecular systems, and data-efficient adaptation to new biological problems.

Projects can be tailored according to the student's background, research interests, and duration of study.

Acute myeloid leukemia (AML) is an aggressive hematologic malignancy driven by clonal expansion of immature myeloid cells in the bone marrow (BM). Remodeling of the BM immune microenvironment contributes to disease progression and therapeutic resistance. However, the cellular composition, functional states, and spatial organization of the human AML BM immune landscape remain poorly defined.

We propose to use ultra-high-plex spatial imaging on FFPE BM specimens from AML patients and healthy controls. A multiplex antibody panel will simultaneously profile major immune and hematopoietic populations along with markers of activation, proliferation, exhaustion, and checkpoint signaling. Single-cell spatial analyses will quantify differences in abundance and phenotype, and evaluate cell–cell proximity, neighborhood composition, clustering, and interactions to identify immunosuppressive niches and preferential localization of dysfunctional effectors near leukemic or regulatory cells.

This study will generate a high-dimensional spatial atlas of the AML immune microenvironment and identify candidate interactions and pathways as biomarkers or targets for immune-based therapies.

Extramedullary multiple myeloma (EMM) is an aggressive and clinically challenging manifestation of multiple myeloma characterized by the dissemination of malignant plasma cells to tissues outside the bone marrow. EMM is associated with poor prognosis, increased treatment resistance, and limited therapeutic options, highlighting the urgent need for novel and effective treatment strategies. This project aims to identify cell surface proteins that represent potential immunotherapeutic targets for EMM using plasma membrane profiling (PMP), a quantitative proteomic approach for comprehensive characterization of the cell surface landscape. PMP will be employed to systematically profile and compare the plasma membrane proteomes of EMM and conventional bone marrow-derived multiple myeloma models. Differentially expressed cell-surface proteins will be identified and prioritized based on their tumour specificity, abundance, and suitability for therapeutic targeting. Candidate targets will subsequently be validated in relevant EMM models and patient samples. The identified targets will provide insights into the molecular characteristics of EMM and may facilitate the development of novel antibody-based, antibody-drug conjugate, or adoptive cellular immunotherapies. Ultimately, this project seeks to identify clinically relevant surface antigens for therapeutic intervention in EMM patients.

Natural Killer/T-cell lymphoma (NKTL) is an aggressive hematological malignancy with limited treatment options and poor outcomes in patients with advanced or relapsed disease. Previous plasma membrane profiling identified a panel of differentially expressed cell surface proteins that may represent potential therapeutic targets in NKTL. This project aims to investigate the biological significance and therapeutic potential of these candidate targets in NKTL. Selected targets will be systematically characterized across NKTL cell lines and patient-derived samples to determine their expression profiles and association with disease characteristics. Functional studies will be performed to investigate the role of these targets in key malignant phenotypes, including cell proliferation, survival and treatment response. Target validation will be conducted before subsequent evaluation of the therapeutic potential of promising candidates using therapeutic antibodies, antibody-based approaches, including antibody-drug conjugates where appropriate and adoptive cellular therapies. These studies will establish whether the identified surface proteins are functionally relevant to NKTL biology and represent actionable therapeutic vulnerabilities. Ultimately, this project aims to translate surfaceome based discoveries into novel therapeutic strategies for NKTL and improve the treatment landscape for patients with this aggressive disease.

Multiple myeloma is a blood cancer caused by the abnormal growth of plasma cells in the bone marrow. MicroRNAs (miRNAs) are small non-coding RNA molecules that regulate gene expression and may influence cancer-cell growth, survival and treatment response. Because miRNAs can also be detected in serum, they have potential as minimally invasive biomarkers for cancer diagnosis and monitoring. Our preliminary profiling of serum samples from patients with multiple myeloma and healthy controls identified a panel of candidate miRNAs with altered circulating levels. However, circulating miRNAs may originate from myeloma cells, blood cells or other components of the bone-marrow microenvironment. Their biological functions in multiple myeloma also remain unclear.
This project will help determine whether serum miRNA changes observed in patients are related to miRNA expression or release by myeloma cells. It will also identify candidate miRNAs that may regulate myeloma-cell proliferation or survival. Promising miRNAs may be investigated further through target-gene analysis, reporter assays and rescue experiments. The results will contribute to our understanding of circulating miRNAs as biomarkers and functional regulators in multiple myeloma.

Lipid nanoparticle (LNP) is a well-developed drug delivery system for mRNA delivery which has been widely used in vaccine development during Covid-19 pandemic. However, delivery efficiency remains a challenge for clinical applications of mRNA. In our project, we developed a series of functional lipid nanoparticles (LNPs) to enhance biological stability of mRNA and its lysosome escape, and thus yielding high translational capacity of mRNA.

In this project, we developed a novel approach for the construction of “3D” mRNA. Controlled interaction between “3D” mRNA and lipid components could induce polyhedral morphology and adjustable rigidness of delivery vehicle, and finally improve the intracellular delivery of mRNA. We propose to test this mRNA delivery platform technology in multiple diseases models, including mRNA vaccine for cancer immunotherapy, crispr/cas9 mRNA system mediated metabolic disease treatment as well as cardiovascular diseases.

Successful T cell immunotherapy for solid tumors requires that T cell can access tumor tissues effectively, which is difficult to achieve. We developed janus-structured nanoparticles modified with targeting ligands which could guide therapeutic T cells to inflammatory tumor microenvironment. We anticipate that this homing system would robustly improve the infiltration of therapeutic Car-T cells to antigen specific solid tumors. Besides, we propose to develop novel mRNA delivery systems based on our proprietary LNP techniques for in-vivo T cell genetic engineering and cancer immunotherapy.

When equipped with wireless technology and image recognition system, deep learning models have been applied in clinic endoscopy examination which allows fast and accurate assistant diagnosis of gastrointestinal diseases. We propose to use a multidisciplinary design principle coupling artificial intelligence techniques and robotic drug delivery system to create a technological interface that allows remote regulation of precision drug delivery. By developing a multitask deep learning algorithm, the endoscopic image information could be analyzed and translated to wireless signal, which will then trigger the ejection of biologic scaffold designed for drug loading to regulate the localization and timing of drug release. We will test this artificial intelligence drug delivery system in multiple gastrointestinal diseases models including bleeding and inflammatory bowel disease.

Cancer is the second leading cause of death in the world. Cancer immunotherapy has brought great hope to cancer patients in recent years, especially with the success in immune checkpoint blockade (ICB) therapy and chimeric antigen receptor (CAR) T-cell therapy. However, their efficacy in solid tumours remains to be improved, which is mainly due to the immunosuppressive tumour microenvironment (TME) that inhibits the infiltration, proliferation, and persistence of cytotoxic T cells. By leveraging advanced nanotechnologies on the delivery of immunomodulators, immunomodulating nanoparticles provide great opportunities to overcome these challenges by switching “cold” tumours to “hot” tumours.
Therefore, we seek to integrate chemistry, biomaterials science, nanomedicine, molecular imaging, immuno-engineering, and oncology fundamentals (1) to design novel nanomedicine-based technologies for effective delivery of immunomodulators; (2) to study the cellular interactions in the cancer-immunity cycle; and (3) to develop new therapies for the treatment of solid tumours.