Skip to main content

Precision medicine for paediatric brain cancer patients

Primary supervisor

David Dowe


  • Ron Firestein

The proposed PhD project aims to build a machine learning/deep learning-based decision support system that provides recommendations on precision medicine for paediatric brain cancer patients based on clinical, genomics and functional dependency data (CRISPR, drug screens).


Brain tumours represent the second most common cancer and the most common solid tumour in childhood in general. Paediatric brain tumour treatments involve surgery, chemotherapy, radiotherapy and steroids. Chemotherapy drugs in particular, by themselves or in combination with other treatments, are very well proven at killing cancer cells, but they also cause severe side effects including significant long-term neurocognitive and neuroendocrine effects in children. A major challenge in chemotherapy-based cancer treatment is predicting clinical response to anti-cancer drugs and determining the precise dosage of drugs on a personalized basis, which precision medicine aims to address. To deliver personalized treatment with high efficacy, identifying molecular disease signatures and matching them with the most effective therapeutic interventions are essential.


The Hudson‐Monash Paediatric Precision Medicine (HMPPM) Program aims to develop and bring genomics‐based medicine to the Victorian paediatric cancer population. Utilising cutting edge genomic technologies combined with sophisticated patient‐derived clinical models, HMPPM will develop and implement personalised cancer treatments, based on the unique molecular footprints that define a patient’s tumour.
HMPPM goes beyond genomic sequencing to identify biomarker coupled targeted therapies that improve treatment efficacy and limit the debilitating side effects of standard chemotherapy and/or radiation therapy.
It is anticipated that this program will lead to substantial improvement in clinical outcome and quality of life (limiting side‐effects) for paediatric solid tumour patients in Victoria, Australia and beyond.
HMPPM Program requires functional dependency data (CRISPR, drug screens) to be integrated with molecular data (generated by HMPPM program, CBTTC Paediatric Brain Tumour Atlas and other sources) for each matched tumour (WGS, RNA‐seq, proteomics, epigenomics) to identify statistically significant associations and map cancer dependencies according to diverse molecular features. Deriving statistically significant features of “omics” data, integrating with functional dependency data to predict statistically significant interactions between genomic profiles and drug sensitives remains very challenging due to size and complexity associated with “omics” data and unrevealed pathway dependencies between “omics” data.


To meet this challenge, the proposed PhD project aims to develop and deploy an AI tool/Decision support system based on machine learning & deep learning. The proposed project aims to (1) build an integrated machine learning/ deep learning framework, (2) facilitate molecular feature selection or dimensionality reduction of the cellular measurements via supervised/unsupervised learning; (3) quantification of drug response; (4) consolidate distinct molecular and clinical features of each patient’s tumour with functional cancer dependency (drug/genetic sensitivity data) (5) Drug response validation via CRISPR data; (6) provide recommendations on precise drug dosages based on genomic and clinical profiles. The project will build machine learning and/or deep learning sub-models and concatenate them to form an integrated scalable framework that supports model evaluation, assess the performance and hyper parameter optimisation based on clinical and molecular data cohorts. The project will also intend to adopt “transfer learning” if any data sparsity issues occur.  It is hoped that the models will be interpretable rather than "black box".




Xia F., Shukla M., Brettin T., Garcia-Cardona C., Cohn J., Allen J., Predicting tumor cell line response to drug pairs with deep learning. BMC Bioinformatics. 2018; 19(Suppl 18): 486. doi: 10.1186/s12859-018-2509-3


Chiu Y., Chen H., Zhang T., Zhang S., Gorthi A., Wang L., Predicting drug response of tumors from integrated genomic profiles by deep neural networks. 2019. BMC Medical Genomics volume 12, Article number: 18


Luo P., Ding Y., Lei X., Wu F., deepDriver: Predicting Cancer Driver Genes Based on Somatic Mutations Using Deep Convolutional Neural Networks. Front Genet. 2019; 10: 13. doi: 10.3389/fgene.2019.00013.


Sakellaropoulos T., Vougas K., Narang S., Koinis F., Kotsinas A., Polyzos A., Moss T., A Deep Learning Framework for Predicting Response to Therapy in Cancer. 2019. Cell Reports. Volume 29, Issue 11, 10 Pages 3367-3373.e4. doi:10.1016/j.celrep.2019.11.017


Yu H., Samuels D., Zhao Y., Guo Y., Architectures and accuracy of artificial neural network for disease classification from omics data. BMC Genomics. 2019; 20: 167. doi: 10.1186/s12864-019-5546-z
Filipp F., Opportunities for Artificial Intelligence in Advancing Precision Medicine. 2019. Current Genetic Medicine Reports volume 7, pages208–213.


Ma Y., Ding Z., Qian Y., Shi X., Castranova V., Harner E., Guo L., Predicting Cancer Drug Response by Proteomic Profiling. Clin Cancer Res. 2006 Aug 1;12(15):4583-9. doi: 10.1158/1078-0432.CCR-06-0290.


IJzendoorn D., Szuhai K., Briaire-de Bruijn I., Kostine M., Kuijjer M., Bovee J., Machine learning analysis of gene
expression data reveals novel diagnostic and prognostic biomarkers and identifies therapeutic targets for
soft tissue sarcomas. PCBI. 2019. Doi: 10.1371/journal.pcbi.1006826


  D. L. Dowe (2020), ``Discussion on the meeting on `Signs and sizes: understanding and replicating statistical findings''', J Royal Statist Soc. (A), vol. 183, no. 2 (February 2020), p453


  Needham, S.L. and D.L. Dowe (2001). Message Length as an Effective Ockham's Razor in Decision Tree Induction. Proc. 8th International Workshop on Artificial Intelligence and Statistics (AI+STATS 2001), pp253-260, Key West, Florida, U.S.A., Jan. 2001


  P. J. Tan and D. L. Dowe (2003). MML Inference of Decision Graphs with Multi-Way Joins and Dynamic Attributes, Proc. 16th Australian Joint Conference on Artificial Intelligence (AI'03), Perth, Australia, 3-5 Dec. 2003, Published in Lecture Notes in Artificial Intelligence (LNAI) 2903, Springer-Verlag, pp269-281


  Wallace, C.S. (2005), ``Statistical and Inductive Inference by Minimum Message Length'', Springer  (Link to the preface [and p vi, also here])


  Wallace, C.S. and D.L. Dowe (1999a). Minimum Message Length and Kolmogorov Complexity, Computer Journal (special issue on Kolmogorov complexity), Vol. 42, No. 4, pp270-283.


Required knowledge

Essential criteria: 

Minimum entry requirements can be found here:

Learn more about minimum entry requirements.


Learn more about minimum entry requirements.