Zur Kurzanzeige

New Computational Tools for Sample Purification and Early-Stage Data Processing in High-Resolution Cryo-Electron Microscopy

dc.contributor.advisorStark, Holger Prof. Dr.
dc.contributor.authorSchulte, Lukas
dc.date.accessioned2019-01-07T10:33:53Z
dc.date.available2019-01-07T10:33:53Z
dc.date.issued2019-01-07
dc.identifier.urihttp://hdl.handle.net/11858/00-1735-0000-002E-E54F-B
dc.identifier.urihttp://dx.doi.org/10.53846/goediss-7220
dc.language.isoengde
dc.rights.urihttp://creativecommons.org/licenses/by-nc-nd/4.0/
dc.subject.ddc571.4de
dc.titleNew Computational Tools for Sample Purification and Early-Stage Data Processing in High-Resolution Cryo-Electron Microscopyde
dc.typedoctoralThesisde
dc.contributor.refereeStark, Holger Prof. Dr.
dc.date.examination2018-09-14
dc.description.abstractengLarge macromolecular protein complexes play an important role in the quest of un- derstanding life at a molecular level. For this endeavor, the highly-resolved 3D structure of a large complex and its changes, which are called conformational dy- namics, are of high interest due to their close relation to biological function. A pop- ular method to gain access to this information is cryo-electron microscopy. Here, biological samples are vitrified in their native buffer conditions and then im- aged in a transmission electron microscope. The large amount of recorded noisy images, called micrographs, are then subjected to a series of sophisticated, compu- tationally expensive algorithmic steps that include particle extraction, averaging of similar particle images, angular assignment and 3D reconstruction. While a growing amount of 3D structures determined by cryo EM reach resolutions that allow atomic model building, a large share of structures never reach that level of possible biolog- ical detail. In this thesis, improvements to two of the problems that can limit the resolution of structures are presented: Sample quality and elimination of suboptimal micro- graphs. First, a new software for simulation of density gradient centrifugation ex- periments, which is a popular method for sample purification, is presented that can aid users in finding the right conditions to optimize their purification proto- cols through new algorithmic approaches. This can lead to purer and more stable samples that improve the recorded images. Then, a new software for live processing and machine learning based classification of TEM images whose quality is judged live by a user is introduced. This software, called the Micrograph Quality Checker, is part of the COW, which is a novel single particle cryo EM image processing framework that covers the whole data analysis workflow. Elimination of unwanted data and collection of metadata in parallel to image acquisition can reduce the data overhead, improve computational speed and can also lead to higher resolution in the obtained structures due to the self-referential image processing.de
dc.contributor.coRefereeFicner, Ralf Prof. Dr.
dc.subject.engcryo-electron microscopyde
dc.subject.engsingle particle imaging frameworkde
dc.subject.engmachine learningde
dc.subject.engdensity gradient centrifugationde
dc.subject.engmicrograph quality criteriade
dc.subject.engsoftwarede
dc.identifier.urnurn:nbn:de:gbv:7-11858/00-1735-0000-002E-E54F-B-4
dc.affiliation.instituteGöttinger Graduiertenschule für Neurowissenschaften, Biophysik und molekulare Biowissenschaften (GGNB)de
dc.subject.gokfullBiologie (PPN619462639)de
dc.identifier.ppn1045789992


Dateien

Thumbnail

Das Dokument erscheint in:

Zur Kurzanzeige