ARCHIVE Snapshot of the Rätsch lab website (Tübingen, then New York) as of February 2016: preserved, not maintained. The group today: bmi.inf.ethz.ch
Personal tools
Home → Rätsch Lab → People → Research Topics → Computational Genome Annotation
 
Document Actions

Computational Genome Annotation

The group’s major contribution to Genomics has been the development of a novel and very accurate gene finding system, called mGene, that uses the latest advances in Machine Learning.

Our system outperformed other gene finding systems in a competition on nematodes (on average over the evaluation criteria for ab initio predictions). We have recently extended this system to additionally take advantage of direct transcriptome measurements, which leads to an improved prediction performance. Moreover, we have developed new techniques that allow us to incorporate and profit from other organisms in order to improve the prediction accuracy, particularly when little is known about the organism to be annotated. In other projects we considered methods further downstream of genome analyses, in particular: (i) methods for the analysis and prediction of alternative splicing events that may lead to several gene products; (ii) the prediction of functional properties of the encoded proteins; and (iii) the analysis of protein-protein interaction data in order to predict functional complexes.

Research Projects

Gene Finding
The group’s major contribution to Genomics has been the development of a novel and very accurate gene finding system, called mGene, that uses the latest advances in Machine Learning. Our system outperformed other gene finding systems in a competition on nematodes (on average over the evaluation criteria). We have recently extended this system to additionally take advantage of direct transcriptome measurements, which leads to an improved prediction performance.
Protein Subcellular Localization
We use "Multiple Kernel Learning" (MKL), an extension to support vector machines (SVMs) to combine and select diverse data sources, to predict the subcellular localization of proteins.
Domain Adaption for Genomic Sequence Analysis
Supervised learning methods are machine learning techniques that use a set of labeled training examples (x_i; y_i) to infer a model that generalizes well to unseen examples x, and therefore allows to make predictions y for these instances. One of the underlying assumptions is that training and test data are drawn from the same distribution. However, in real-world applications, and specifically in computational biology, one often faces the situation that little or no training data is available for the problem of interest (the target domain), whereas a large corpus of labeled data exists for a related but different problem (the source domain). While it is highly desirable to exploit the labeled source data points to improve performance on the target domain, it clearly violates the basic assumptions of machine learning. Methods to tackle this so-called domain adaptation problem have recently caught increasing attention in the machine learning community and are subject of consideration in this work. In particularly we examine their effectiveness for the problem of splice site prediction across different organisms with varying evolutionary distance.
Inference of Alternative Splicing
Transcription Start and Stop Detection
We develop new methods for finding transcription start sites (TSS) of RNA Polymerase II binding genes in genomic DNA sequences. Employing Support Vector Machines with advanced sequence kernels, we achieve drastically higher prediction accuracies than state-of-the-art methods.

Publications