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.

Currently, I am persuing my Ph.D. with Dr Viji M. Draviam at School of Biological and Chemical Sciences and Dr Nishanth Sastry in the Department of Informatics at King's College 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.


Queen Mary University London
School Biological and Chemical Sciences
King's College London
Department of Informatics
Ph.D. candidate
Sep 2016 - present (United Kingdom)

My research focus is mainly on modelling of various cell compartment dynamics in human cancer cells.

  • To conduct my research, I use a deconvolution microscopy that allows me to image living cells through time and space. In particular, I am interested in a bulky sub-cellular structure called spindle. The two major functions of the spindle is to orient and position itself correctly which is a crucial process to define the plane of cell division and tissue development. Having been trained in cell culture techniques, I perform various experiments under different drug-treated conditions. Then, I use a DeltaVision microscope system to image hundreds of cells in 4D.
  • Simultaneously, I am developing my own spinX Software using image processing technologies from Computer Vision and Deep Learning with the goal to automatically analyze those microscopy movies by precisely segmenting different cell compartments and tracking them through time and space.
  • With the extracted information from the movies, I start building a statistical framework to find patterns within the data structure and to draw biological meaningful conclusions to shed light on how spindle movements are governed by cells.
University of Bremen
Department of Mathematics and Computer Sciences
M.Sc. Biostatistics
2014 - 2016 (Germany)

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.

  • During my Master’s degree, I have had the opportunity to delve into the intricacies of statistics and hone the mathematical ideas of the quantitative assumptions that underlie proper techniques in modelling. Major Courses in: Complex Statistical Modelling, Statistical Programming and Data Mining, non-parametric Methods, Survival analysis, Time-series analysis, Hypothesis testing. Non-mathematical courses taken in: Clinical Trials, Basic Medicine, Oncology, Ethics.
University of Tuebingen
Department Economics and Social Sciences
B.A. Social Sciences
2011 - 2014 (Germany)

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.

  • With a strong interests in Statistics, I visited most courses in Econometry and Data management where I got first insights in Probability Theories and its application in social and economic data. My minor degree was in Philosophy where I have visited lectures on logic, pratical philosophy, philosophy of law and science.

Work experiences

University College London
Department of Mechanical Engineering
Queen Mary University
School of Medicine and Dentistry
Software Developer
Feb 2017 - Jul 2017 (United Kingdom)

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.

Leibniz Institute BIPS GmbH
Division of Statistics
Research student
Jan 2016 - September 2017 (Germany)

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.

Competence Centre for Clinical Trials
Division of Statistics
Aug 2015 - Sept 2015 (Germany)

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.

HTML Design AG
Division of Online Marketing
Aug 2013 - Dec 2013 (Germany)

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).


MARK2/Par1b present at retraction fibres corrects spindle off-centering induced by actin disassembly (2018)
Madeleine Hart, Ihsan Zulkipli, Roshan Lal Shrestha, David Dang, Duccio Conti, Izabela Kujawiak & Viji M. Draviam (2018).

Available here: [pdf]

Abstract: Tissue maintenance requires adequate cell proliferation and a directed plane of cell division. While it is clear that the division-plane can be determined by retraction fibres that direct spindle movements, the components of retraction fibres that direct spindle movements are poorly understood. We report MARK2/Par1b kinase as a novel component of actin-rich retraction fibres, important for directed spindle movements. A kinase-dead mutant of MARK2 reveals the ability of MARK2 to monitor actin status. MARK2 localisation at actin-rich retraction fibres, but not the rest of the cortical membrane or centrosome, is dependent on its activity, highlighting a specialised spatial regulation of MARK2. By subtly perturbing the actin cytoskeleton, we demonstrate MARK2 has a role in correcting spindle off-centering, induced by actin disassembly. In addition to this mitotic role, we show MARK2 has a post-mitotic role in ensuring normal G1-S transition and cell proliferation. We propose that MARK2 provides a molecular framework to integrate cortical signals and cytoskeletal changes in both mitosis and interphase.

Spindle rotation in human cells is reliant on a MARK2-mediated equatorial spindle-centering mechanism (2018)
Ihsan Zulkipli, Joanna Clark, Madeleine Hart, Roshan L. Shrestha, Parveen Gul, David Dang, Tami Kasichiwin, Izabela Kujawiak, Nishanth Sastry & Viji M. Draviam.

Available here: [pdf]

Abstract: The plane of cell division is defined by the final position of the mitotic spindle. The spindle is pulled and rotated to the correct position by cortical dynein. However, it is unclear how the spindle’s rotational center is maintained and what the consequences of an equatorially off centered spindle are in human cells. We analyzed spindle movements in 100s of cells exposed to protein depletions or drug treatments and uncovered a novel role for MARK2 in maintaining the spindle at the cell’s geometric center. Following MARK2 depletion, spindles glide along the cell cortex, leading to a failure in identifying the correct division plane. Surprisingly, spindle off centering in MARK2-depleted cells is not caused by excessive pull by dynein. We show that MARK2 modulates mitotic microtubule growth and length and that codepleting mitotic centromere-associated protein (MCAK), a microtubule destabilizer, rescues spindle off centering in MARK2-depleted cells. Thus, we provide the first insight into a spindle-centering mechanism needed for proper spindle rotation and, in turn, the correct division plane in human cells.

Thanks to my sponsors