ESR12 – Rohit Thakur – based in Leeds

Project Title: Developing statistical and bioinformatic analysis of genomic data from tumours
PhD Awarded: January 2019

Deploying the concepts from mathematics and applying to understand biological pathways and expression networks is an interesting concept. This interest grew in me while pursuing the undergraduate course in Bioinformatics at VIT University, India. An aspiring computational biologist also require interdisciplinary knowledge of biochemical concepts as well as, needs to be adept in computational and programming proficiency. Research experiences at Alpha Net Technologies (India), IISER Pune (India) and EMBL-EBI (UK), provided me with the opportunity to acquire this proficiency.

I always look forward to become part of intellectual discussion groups and had the experience of working for an initiative called RSG India that promotes networking amongst computational biologists in India. I am an active sports player and play cricket and basketball. I also seek to help humanity and have been a part of the NGO named “MAD” (Make A Difference) helping underprivileged students by guiding them to pursue their dream. Experiences inside and outside the lab helped me recognize my ‘driving force’ and career aspirations. My research interest lies broadly in the area of cancer genomics. I look forward to a career as a researcher developing algorithms and designing virtual diagnostic systems which would impact healthcare treatments in the long run.

Research Summary

During my PhD, I developed prognostic gene signatures using genomic datasets from one of the largest primary melanoma cohorts to date. Using unsupervised classification approach, I devised a transcriptome-based molecular signature which predicts prognosis, particularly in stage I one tumors. This is an important contribution to the field as 90% of the melanomas are diagnosed early and this signature could be useful in identifying patients who may benefit from receiving early adjuvant therapies. Furthermore, this molecular signature also demonstrated predictive value in identifying patients with metastatic melanoma treated with immunotherapy who are not likely to respond and this work has now been published in Clinical Cancer Research.

During my PhD, I also received training in machine learning and developed classification models that predict high- and low-risk of melanoma relapse after initial diagnosis. The MELGEN program has been instrumental in defining my career path by providing numerous training opportunities to improve my technical and communication skills as well as scientific interactions with world leaders in melanoma research. As part of the MELGEN program, I conducted short research visits to Dr. Goran Jonsson’s group at University of Lund, Sweden, Dr. Will Spooner at Eagle Genomics, Cambridge, UK and Dr. Manolis Kellis’ group at Massachusetts Institute of Technology, USA where I received training in differential gene expression analysis and machine learning using genomic datasets. As a result of this training I was successful in developing a machine learning model that in future would be helpful in stratifying patients into low risk and high risk of relapse and subsequently would inform clinical decisions.


Thakur R, Laye JP, Lauss M, Diaz JMS, O’Shea SJ, Poźniak J, Filia A, Harland M, Gascoyne J, Randerson-Moor JA, Chan M, Mell T, Jönsson G, Bishop DT, Newton-Bishop J, Barrett JH. Transcriptomic Analysis Reveals Prognostic Molecular Signatures of Stage I Melanoma. Clin Cancer Res. 2019 Dec 15;25(24):7424-7435. Available at: