OxCSML : People : Tammo Rukat

GitHub

Matrix Factorisation, Bayesian Neural Nets, Deep Generative Models, Statistical Genetics

I am a student with Chris Holmes and Christopher Yau in the Department of Statistics and at the Wellcome Trust Centre for Human Genetics at the University of Oxford.

I am doing my doctorate within the framework of the Systems Approaches to Biomedical Sciences Centre for Doctoral Training.

Publications

2017

T. Rukat , C. C. Holmes , M. K. Titsias , C. Yau , Bayesian Boolean Matrix Factorisation, arXiv preprint arXiv:1702.06166, 2017.

@article{rukat2017bayesian,
  title = {Bayesian Boolean Matrix Factorisation},
  author = {Rukat, Tammo and Holmes, Chris C and Titsias, Michalis K and Yau, Christopher},
  journal = {arXiv preprint arXiv:1702.06166},
  year = {2017}
}

T. Rukat , C. C. Holmes , M. K. Titsias , C. Yau , Bayesian Boolean Matrix Factorisation, 2017.
Boolean matrix factorisation aims to decompose a binary data matrix into an approximate Boolean product of two low rank, binary matrices: one containing meaningful patterns, the other quantifying how the observations can be expressed as a combination of these patterns. We introduce the OrMachine, a probabilistic generative model for Boolean matrix factorisation and derive a Metropolised Gibbs sampler that facilitates efficient parallel posterior inference. On real world and simulated data, our method outperforms all currently existing approaches for Boolean matrix factorisation and completion. This is the first method to provide full posterior inference for Boolean Matrix factorisation which is relevant in applications, e.g. for controlling false positive rates in collaborative filtering and, crucially, improves the interpretability of the inferred patterns. The proposed algorithm scales to large datasets as we demonstrate by analysing single cell gene expression data in 1.3 million mouse brain cells across 11 thousand genes on commodity hardware.
```
@article{rukat2017_bayes-boolean,
  title = {Bayesian Boolean Matrix Factorisation},
  author = {Rukat, Tammo and Holmes, Chris C. and Titsias, Michalis K. and Yau, Christopher},
  archiveprefix = {arXiv},
  year = {2017},
  eprint = {1702.06166},
  primaryclass = {stat.ML}
}
```

2015

T. Rukat , A. Baker , A. Quinn , M. Woolrich , Resting state brain networks from EEG: Comparing hidden Markov states with classical microstates, Proceedings of the 5th NIPS Workshop on Machine Learning and Interpretation in Neuroimaging (MLINI), Dec. 2015.

@article{Rukat2015d,
  title = {Resting state brain networks from EEG: Comparing hidden Markov states with classical microstates},
  author = {Rukat, T. and Baker, A. and Quinn, A. and Woolrich, M.},
  journal = {Proceedings of the 5th NIPS Workshop on Machine Learning and Interpretation in Neuroimaging (MLINI)},
  year = {2015},
  month = dec
}

T. Rukat , Distributed analysis of expression quantitative trait loci in Apache Spark, Jul-2015.
A major challenge in current genomic research is the development of computational and statistical tools that are capable of analysing the ever increasing amount of data provided by next generation sequencing methods. Here we investigate the potential of the open-source distributed computing framework Apache Spark, to facilitate fast, horizontally scalable genetic data analysis, exemplified by the search for expression quantitative trait loci (eQTL) in the 1000 Genomes dataset. An algorithm for trans eQTL analysis is proposed and enables the analysis of all available pairwise correlations of gene-SNP- pairs in the dataset (> 3 × 10 11 ) in approximately 90 minutes on a cluster with 768 worker nodes. The algorithm scales linearly with the problem size, and benefits from larger cluster sizes up to 768 nodes. A second algorithm for cis eQTL analysis is introduced but shown to be inherently difficult to design within a distributed system and therefore only recommended for very targeted investigations. Though constantly enhanced and extended, Spark already offers a simple and powerful environment, suitable for distributed genetic data analysis.
```
@unpublished{rukat15:_distr_apach_spark,
  title = {Distributed analysis of expression quantitative trait loci in Apache Spark},
  author = {Rukat, Tammo},
  month = jul,
  year = {2015},
  comment = {https://github.com/TammoR/spark_eqtl},
  file = {Rukat_Spark_eQTL.pdf:Rukat_Spark_eQTL.pdf:PDF},
  owner = {tammo},
  timestamp = {2015.08.10}
}
```

T. Rukat , S. A. Reinsberg , A Bayesian Information Criterion for Multi-Model Inference in DCE-MRI, in Proc. Intl. Soc. Mag. Reson. Med. 23 (2015) 2339, 2015, no. 2339.

@inproceedings{Rukat2015a,
  title = {A {Bayes}ian Information Criterion for Multi-Model Inference in {DCE-MRI}},
  author = {Rukat, Tammo and Reinsberg, Stefan A.},
  booktitle = {Proc. Intl. Soc. Mag. Reson. Med. 23 (2015) 2339},
  year = {2015},
  number = {2339},
  file = {Rukat2014a_Information Criteria weighted Parameter Estimates in DCE-MRI.pdf:Rukat2015.pdf:PDF;Rukat2014a_Information Criteria weighted Parameter Estimates in DCE-MRI_Abstract.pdf:Rukat2014a_Information Criteria weighted Parameter Estimates in DCE-MRI_Abstract.pdf:PDF},
  owner = {tammo},
  timestamp = {2014.09.18}
}

T. Rukat , S. Walker-Samuel , S. A. Reinsberg , Dynamic contrast-enhanced MRI in mice: An investigation of model parameter uncertainties, Magnetic Resonance in Medicine, vol. 73, 1979–1987, 2015.
Purpose To establish the experimental factors that dominate the uncertainty of hemodynamic parameters in commonly used pharmacokinetic models. Methods By fitting simulation results from a multiregion tissue exchange model (Multiple path, Multiple tracer, Indicator Dilution, 4 region), the precision and accuracy of hemodynamic parameters in dynamic contrast-enhanced MRI with four tracer kinetic models is investigated. The impact of various injection rates as well as imprecise knowledge of the arterial input functions is examined. Results Fast injections are beneficial for Ktrans precision within the extended Tofts model and within the two-compartment exchange model but do not affect the other models under investigation. Biases from errors in the arterial input functions are mostly consistent in size and direction for the simple and the extended Tofts model, while they are hardly predictable for the other models. Errors in the hematocrit introduce the greatest loss in parameter accuracy, amounting to an average Ktrans bias of 40% for a 30% overestimation throughout all models. Conclusion This simulation study allows the detailed inspection of the isolated impact from various experimental conditions on parameter uncertainty. Because parameter uncertainty comparable to human studies was found, this study represents a validation of preclinical dynamic contrast-enhanced MRI for modeling human tumor physiology. Magn Reson Med, 2014. Â\copyright 2014 Wiley Periodicals, Inc.
```
@article{Rukat2015,
  title = {Dynamic contrast-enhanced {MRI} in mice: An investigation of model parameter uncertainties},
  author = {Rukat, T. and Walker-Samuel, S and Reinsberg, S. A.},
  journal = {Magnetic Resonance in Medicine},
  year = {2015},
  pages = {1979–1987},
  volume = {73},
  copyright = {Â© 2014 Wiley Periodicals, Inc.},
  doi = {10.1002/mrm.25319},
  file = {Rukat2014_Dynamic contrast-enhanced {MRI} in mice\: An investigation of model parameter uncertainties.pdf:Rukat2014.pdf:PDF},
  issn = {1522-2594},
  keywords = {Dynamic contrast-enhanced-{MRI}; Perfusion; Permeability; pharmacokinetic models; tracer-kinetic models},
  language = {en},
  shorttitle = {Dynamic contrast-enhanced {MRI} in mice},
  urldate = {2014-07-27}
}
```

2014

T. Rukat , S. A. Reinsberg , Information Criteria weighted Parameter Estimates in DCE-MRI, in Proc. Intl. Soc. Mag. Reson. Med. 22 (2014) 2741, 2014, no. 2741.

@inproceedings{Rukat2014,
  title = {Information Criteria weighted Parameter Estimates in {DC}E-{MRI}},
  author = {Rukat, Tammo and Reinsberg, Stefan A.},
  booktitle = {Proc. Intl. Soc. Mag. Reson. Med. 22 (2014) 2741},
  year = {2014},
  number = {2741},
  file = {Rukat2014a_Information Criteria weighted Parameter Estimates in DCE-MRI.pdf:Rukat2014a.pdf:PDF;Rukat2014a_Information Criteria weighted Parameter Estimates in DCE-MRI_Abstract.pdf:Rukat2014a_Information Criteria weighted Parameter Estimates in DCE-MRI_Abstract.pdf:PDF},
  owner = {tammo},
  timestamp = {2014.09.18}
}

2013

T. Rukat , Parameter Uncertainties in Tracer Kinetic Modelling of Dynamic Contrast Enhanced MRI, Master's thesis, Humboldt University Berlin / University of British Columbia, 2013.

@mastersthesis{Rukat2013a,
  title = {Parameter Uncertainties in Tracer Kinetic Modelling of Dynamic Contrast Enhanced {MRI}},
  author = {Rukat, Tammo},
  school = {Humboldt University Berlin / University of British Columbia},
  year = {2013},
  month = oct,
  file = {Rukat2013a_Parameter Uncertainties in Tracer Kinetic Modelling of Dynamic Contrast Enhanced MRI.pdf:Rukat2013a.pdf:PDF},
  owner = {tammo},
  timestamp = {2014.09.18}
}

T. Rukat , S. Walker-Samuel , S. A. Reinsberg , AIF Induced Limits of Parameter Uncertainty in Pharmakokinetic Models of Pre-Clinical DCE-MRI, in Proc. Intl. Soc. Mag. Reson. Med. 21 (2013) 2214, 2013, no. 2214.

@inproceedings{Rukat2013,
  title = {{AIF} Induced Limits of Parameter Uncertainty in Pharmakokinetic Models of Pre-Clinical {DCE-MRI}},
  author = {Rukat, T. and Walker-Samuel, S. and Reinsberg, S. A.},
  booktitle = {Proc. Intl. Soc. Mag. Reson. Med. 21 (2013) 2214},
  year = {2013},
  number = {2214},
  file = {Rukat2013_AIF Induced Limist of Parameter Uncertainty in Pharmakokinetic Models of Pre-Clinical DCE-MRI.pdf:Rukat2013.pdf:PDF;Rukat2013_AIF Induced Limist of Parameter Uncertainty in Pharmakokinetic Models of Pre-Clinical DCE-MRI_Abstract.pdf:Rukat2013_AIF Induced Limist of Parameter Uncertainty in Pharmakokinetic Models of Pre-Clinical DCE-MRI_Abstract.pdf:PDF},
  journal = {Proc. Intl. Soc. Mag. Reson. Med. 21 (2013) 2214}
}