I am a Ph.D. student in the BBSRC - London Interdisciplinary Doctoral Programme Consortium in London, United Kingdom. I worked with Dr Paul Fromme in the Department of Mechanical Engineering at University College London and Dr Stéphanie Kermorgant at Barts and Medical and Dentistry at Queen Mary University of London.
Having been trained as a Statistician and Computer Scientist prior to coming to London, now I see myself as someone who combines Statistics, Deep Learning in Computer vision and Cell Biology together to seek advances to conduct research on cancer from a quantitative view.
My research focus is mainly on modelling of various cell compartment dynamics in human cancer cells.
My master thesis was on Detection of genomic clusters in network graphs by a correlation-based Markov Cluster Algorithm. The nature of graph theory can be mathematically represented by the adjacency matrix. However, in genomic data links (edges) between genes (vertices) are unknown. A direct application with any graph-based algorithm is not possible. I solved this problem with a statistical approach by estimating the edges using partial correlations and could demonstrate that this optimization step can outperform other traditional unsupervised Machine Learning algorithms in a simulation study as well as on 36 real genomic data sets in terms of cluster recovery accuracy.
In partial fulfilment of my bachelor degree, I investigated the impact of economic growth on income inequality. The income and wealth distribution of a nation can be measured by the GINI-coefficient. Mathematically speaking, it is a normalized ratio based on the Lorenz curve which plots the proportion of the the total income of a population against the cumulatate earnings. I used these messurements to analyse income inequality among the OECD countries.
The cell–surface receptor Met is a receptor tyrosine kinase (RTK) that, after binding with its only known ligand, hepatocyte growth factor (HGF), induces kinase catalytic activity. This, in turn, triggers transphosphorylation of Y1234 and Y1235 initializing a spectrum of signaling cascades which lead to invasive growth that is important during embryonic development. Dysregulated Met via mutation or overexpression is often a cause for the development of various human cancer. In collaboration with Dr. Paul Fromme at UCL and Dr. Stéphanie Kermorgant QMUL, I developed the metTracker Software based on methodologies from Computer Vision to quantify and model Met particle trafficking in pancreatic cancer cells.
Employed in the field of Statistics working on Big Data with the goal to optimize the graph-theory based Markov-cluster Algorithm outperform traditional unsupervised Machine Learning algorithms: Common algorithms, such as hierarchical clustering or k-means, lack sufficiency especially when applied to genomic data. Moreover, in larger networks the number of clusters are mostly unpredictable, whereas the number of cluster is one of the requirements for e.g. k-means algorithm. To solve this problem, I formulated and extended the graph-based Markov-Cluster Algorithm by using partial correlation matrices to predict associations between genes in a gene expression network and demonstrated that by using this strategy the cluster structure in simulated and real data can be recovered with higher accuracy compared to traditional methods.
As part of the Statistics division, I participated in ”ExMo”, a trial with the goal to improve health, and actively helped shape in statistical data analysis as well as in modelling and programming macros for automation processes with SAS.
Nowadays, it is imperative for every company to have a website. It helps the company to establish credibility as a business. Users trust search engines like Google or Bing and achieving one of those rare top spots in search engine ranking lists signals to searchers that your site is a credible source which in turn leads to more clicks and traffic your site will generate. Multiple strategies, best practices, and actions are necessary to improve the visibility of a website and to maintain the position in the search results. As part of the Online Marketing team, I worked closely with web developers to implement measures for search engine optimization (SEO) and to analyze and improve search engine marketing campaigns (SEM).