MASTER’S AND PHD STUDENT POSITIONS
Master’s and PhD student positions are available in Biomedical Computing
Group led by Professor Jagath Rajapakse (https://personal.ntu.edu.sg/asjagath/)
at Nanyang Technological University, Singapore. Interested applicants should
forward their CVs and academic transcripts to Professor Jagath Rajapakse at
asjagath@ntu.edu.sg.
Topics currently available for research:
Functional and structural connectome analysis:
1. Encoding the
connectome: Encoding involves learning
representations from functional MRI and diffusion tensor imaging (DTI) scans to
disease classification and predict clinical measures and drug response. Deep
learning techniques are most popular though such techniques require large
neuroimaging datasets. In this project, we will develop graph neural network
approaches to preserve the graphical structure of the connectome and
unsupervised learning technique to overcome the paucity of neuroimaging data.
2. Decoding the
connectome: By decoding the connectome we identify
brain regions and connections that are implicated in brain disease. This
requires selection of features and interpreting functioning therein of deep
neural networks. In this project, we will develop deep neural networks such as
graph neural networks and identifying salient nodes therein detecting brain
disease.
3. Combining
neuroimaging data with multi-omics data: To investigate
neurodegenerative disease, both neuroimaging and multi-omics data are
increasingly being gathered. However, techniques to integrate neuroimaging and
omics data are lacking and we will investigate deep learning approaches that
integrate multi-omics and neuroimaging data to detect brain disease and
identify potential drug targets.
Multi-omics data analysis:
1. Drug response
prediction from multi-omics data: We want to develop
methods to predict drug response from multi-omics (genomics, proteomics,
transcriptomics, and epigenomics) data gathered from patients. We will first
develop deep neural network approaches for drug response prediction from
multi-omics data gathered from cell lines. Thereafter, we will develop
translational deep learning methods to translate such models on patient
samples.
2. Predicting
drug-target interactions from drug response data: We will develop deep
neural networks approaches to predict drug response by combining both drug
chemical data and multi-omics data. We will use graph neural network approaches
with attention mechanisms and use salience backpropagation techniques to identify
chemical features and omics features that are responsible and interacting with
one another for creating a drug response.
Explainable and Generative AI:
Notes to potential candidates:
· NTU requires minimum
GRE and TOEFL scores for admission to graduate programs
· Students with
excellent academic scores and prior research experience are eligible for
scholarships
· There are research
assistantship positions for self-funding master’s and PhD students, which will
cover most of their expenses
· Visiting students
(for at least for two semesters) are considered on a case-by-case basis
· All interested
students should send their academic transcripts and CV to Professor Jagath
Rajapakse (asjagath@ntu.edu.sg).