FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W86631861> ?p ?o ?g. }

• W86631861 abstract "The property of membrane organisation captures the relationship of a protein to cellular membranes and is defined by the presence of signal peptides and transmembrane domains in the protein sequence. The membrane organisation of a protein will influence the way the protein is sorted by the cell, and constrain its eventual subcellular distribution. I have developed a computational pipeline, MemO, to automatically annotate the membrane organisation of entire proteomes from eukaryotic organisms, using high-confidence consensus algorithms to generate annotation of membrane, secreted and soluble proteins. In the pipeline, eight independent methods are used to generate raw data for the consensus algorithms, with specific attention to addressing issues of conflict between predicted features. The use of consensus methods rather than single prediction methods results in a highly reliable prediction pipeline that has been designed to generate biologically useful annotation. The MemO pipeline has been applied in a range of biological contexts, including functional genomics, developmental biology, cell biology, computational cell biology and comparative genomics. Specifically, three applications of the pipeline are explored, describing my use of MemO as an annotation tool, as a technique for target selection from expression array analysis, and as a tool to compare the properties of eukaryotic proteomes. Application of MemO to the mouse proteome, as recently defined through the sequencing of the mammalian transcriptome, lead to the discovery of considerable diversity in the use of signal peptides and helical transmembrane domains. It was observed that variation in transcripts, caused by variation in transcriptional start and stop locations and alternative splicing, resulted in the variable use of these features, with often dramatic impact on the resulting membrane organisation and subcellular location of the protein products. As a result of these observations, I analysed the variable use of genomic regions encoding these features in a set of >8000 transcriptional units where two or more protein products were produced. I discovered that 39% of these transcriptional units containing predicted signal peptides also demonstrated differential use of this feature, while two thirds of transcriptional units that also contained predicted transmembrane domains either used variable genomic regions to encode transmembrane domains, or varied the number of transmembrane domains present, while 35% also encode a soluble protein. As multiple prediction methods are increasingly marshalled to provide insight into the function of unknown proteins, including the use of methods to detect protein domains and family membership, conflicts between predictions can sometimes occur. The observation that some domains known to be soluble contained predicted transmembrane domains led me to explore the extent of this problem. I discovered a set of soluble protein domains that confound transmembrane domain prediction methods and cause systematic errors in the annotation of membrane proteins across whole proteomes. I developed a method for resolving these contradictions based on the conflict profiles of the competing predictions, and established guidelines for resolving conflicts within more structurally diverse protein families. The result of this analysis provides a strategy for improving membrane protein prediction by identifying false positive predictions in domains that do not span the membrane. In conclusion, this work focused initially on the development of the MemO pipeline, which now provides a method for generating high confidence membrane organisation annotation in very large eukaryotic data sets. Then, the differential use of signal peptides and transmembrane domains in the variable protein products of transcriptional units was explored and characterised, and conflicting predictions between transmembrane domains and protein functional domains and families were explored and resolved. This work highlights the importance of predicting multiple features and resolving the resulting conflicts to characterise protein function, as the prediction of single features in isolation can be misleading. The work also highlights a surprising diversity in membrane organisation of proteins products and the extent of variable use of genomically encoded features in the mammalian proteome." @default.
• W86631861 created "2016-06-24" @default.
• W86631861 creator A5018252802 @default.
• W86631861 date "2023-10-16" @default.
• W86631861 modified "2023-10-17" @default.
• W86631861 title "Defining the Membrane Organisation of Eukaryotic Proteins" @default.
• W86631861 doi "https://doi.org/10.14264/158481" @default.
• W86631861 hasPublicationYear "2023" @default.
• W86631861 type Work @default.
• W86631861 sameAs 86631861 @default.
• W86631861 citedByCount "0" @default.
• W86631861 crossrefType "dissertation" @default.
• W86631861 hasAuthorship W86631861A5018252802 @default.
• W86631861 hasConcept C41008148 @default.
• W86631861 hasConcept C70721500 @default.
• W86631861 hasConcept C86803240 @default.
• W86631861 hasConcept C95444343 @default.
• W86631861 hasConceptScore W86631861C41008148 @default.
• W86631861 hasConceptScore W86631861C70721500 @default.
• W86631861 hasConceptScore W86631861C86803240 @default.
• W86631861 hasConceptScore W86631861C95444343 @default.
• W86631861 hasLocation W866318611 @default.
• W86631861 hasOpenAccess W86631861 @default.
• W86631861 hasPrimaryLocation W866318611 @default.
• W86631861 hasRelatedWork W1596801655 @default.
• W86631861 hasRelatedWork W2130043461 @default.
• W86631861 hasRelatedWork W2350741829 @default.
• W86631861 hasRelatedWork W2358668433 @default.
• W86631861 hasRelatedWork W2376932109 @default.
• W86631861 hasRelatedWork W2382290278 @default.
• W86631861 hasRelatedWork W2390279801 @default.
• W86631861 hasRelatedWork W2748952813 @default.
• W86631861 hasRelatedWork W2899084033 @default.
• W86631861 hasRelatedWork W2530322880 @default.
• W86631861 isParatext "false" @default.
• W86631861 isRetracted "false" @default.
• W86631861 magId "86631861" @default.
• W86631861 workType "dissertation" @default.