Direkt zum InhaltDirekt zur SucheDirekt zur Navigation
▼ Zielgruppen ▼

Humboldt-Universität zu Berlin - IRI Life Sciences

Humboldt-Universität zu Berlin | IRI Life Sciences | Scientific Events | Colloquium | Machine Learning | IRI-Colloquium: Marcel Schulz - Learning gene regulation from large epigenomics and transcriptomics data

IRI-Colloquium: Marcel Schulz - Learning gene regulation from large epigenomics and transcriptomics data

Abstract: Improving our understanding of gene regulation in the context of development and diseases, such as cancer, is one of promises of the large data collection efforts currently underway. However, the complexity of gene regulation makes it anything but trivial to condense information from large amounts of sequencing data into testable biological hypotheses. I will present two novel computational approaches toward that goal. First, I will describe a novel machine learning approach for learning, possibly distal, transcriptional regulatory elements for gene expression from paired epigenomics and transcriptomics data. I will delineate the mathematical formulation and our experimental results for their validation. Second, I will present a multitasking-based optimization approach that can be used to learn regulatory associations for regulators, such as microRNAs and transcription factors, on the level of individual transcript isoforms. I will present results that illustrate that transcript-isoform specific (mis)regulation may be an important regulatory switch in cancer cells.
  • IRI-Colloquium: Marcel Schulz - Learning gene regulation from large epigenomics and transcriptomics data
  • 2019-10-24T16:00:00+02:00
  • 2019-10-24T17:00:00+02:00
  • Abstract: Improving our understanding of gene regulation in the context of development and diseases, such as cancer, is one of promises of the large data collection efforts currently underway. However, the complexity of gene regulation makes it anything but trivial to condense information from large amounts of sequencing data into testable biological hypotheses. I will present two novel computational approaches toward that goal. First, I will describe a novel machine learning approach for learning, possibly distal, transcriptional regulatory elements for gene expression from paired epigenomics and transcriptomics data. I will delineate the mathematical formulation and our experimental results for their validation. Second, I will present a multitasking-based optimization approach that can be used to learn regulatory associations for regulators, such as microRNAs and transcription factors, on the level of individual transcript isoforms. I will present results that illustrate that transcript-isoform specific (mis)regulation may be an important regulatory switch in cancer cells.
What
  • Colloquium
  • Upcoming
  • machine learning
When Oct 24, 2019 from 04:00 PM to 05:00 PM (Europe/Vienna / UTC200) iCal
Where IRI Life Sciences, Philippstr. 13, Building 18, 3rd Floor, Room 410
Contact Name
Contact Phone 30209347904

Poster Bild