ORCID Profile
0000-0002-5804-6031
Current Organisations
Universität Innsbruck
,
Quantium
Does something not look right? The information on this page has been harvested from data sources that may not be up to date. We continue to work with information providers to improve coverage and quality. To report an issue, use the Feedback Form.
Publisher: MDPI AG
Date: 02-05-2023
DOI: 10.3390/INFO14050271
Abstract: Quantifying the dissimilarity of two texts is an important aspect of a number of natural language processing tasks, including semantic information retrieval, topic classification, and document clustering. In this paper, we compared the properties and performance of different dissimilarity measures D using three different representations of texts—vocabularies, word frequency distributions, and vector embeddings—and three simple tasks—clustering texts by author, subject, and time period. Using the Project Gutenberg database, we found that the generalised Jensen–Shannon ergence applied to word frequencies performed strongly across all tasks, that D’s based on vector embedding representations led to stronger performance for smaller texts, and that the optimal choice of approach was ultimately task-dependent. We also investigated, both analytically and numerically, the behaviour of the different D’s when the two texts varied in length by a factor h. We demonstrated that the (natural) estimator of the Jaccard distance between vocabularies was inconsistent and computed explicitly the h-dependency of the bias of the estimator of the generalised Jensen–Shannon ergence applied to word frequencies. We also found numerically that the Jensen–Shannon ergence and embedding-based approaches were robust to changes in h, while the Jaccard distance was not.
Publisher: American Thoracic Society
Date: 12-2018
Publisher: Elsevier BV
Date: 02-2016
DOI: 10.1016/J.JPROT.2015.12.015
Abstract: Many diseases are associated with protein species perturbations. A prominent ex le of an established diagnostic marker is the glycated protein species of hemoglobin, termed HbA1c. HbA1c concentration is increased in the blood of diabetes mellitus patients due to their poor control of blood glucose levels resulting in an increased non-enzymatic glycosylation of hemoglobin producing HbA1c. This important diagnostic marker is routinely measured in the blood of diabetes patients. As in the case of HbA1c, protein species can mirror pathophysiological events. Shifts in the levels of protein species can be associated with or even be responsible for disease making them well suited as diagnostic markers. However, only a few protein species are currently used as diagnostic markers in routine clinical chemistry laboratories, despite being widely established in clinical proteomics research. This review provides an overview of the biochemical characteristics associated with protein species as well as ex les of pathophysiological mechanisms, which cause modifications in the protein species composition, thereby emphasizing the importance of screening for protein markers at the species level. Further, we highlight techniques, which are currently utilized for investigating protein species markers in clinical research. The success rate of FDA approved diagnostic protein markers until today is very low compared to the number of published candidate disease markers. It is hypothesized that one important reason is the gene-centric view which is still followed in clinical proteomics: In many investigations proteins are still digested in small peptides thus making it nearly impossible to discriminate between healthy proteins and pathologic proteins causing diseases. Thus this review is focusing on the biochemistry and patho-biochemistry of proteins, is highlighting the need for screening for disease markers on the protein species level and is giving an overview about available techniques.
Publisher: American Chemical Society (ACS)
Date: 17-07-2018
Publisher: Elsevier BV
Date: 02-2016
Publisher: Impact Journals, LLC
Date: 10-11-2017
Publisher: Public Library of Science (PLoS)
Date: 10-06-2021
DOI: 10.1371/JOURNAL.PONE.0253084
Abstract: Rickettsioses are neglected and emerging potentially fatal febrile diseases that are caused by obligate intracellular bacteria, rickettsiae. Rickettsia ( R .) typhi and R . prowazekii constitute the typhus group (TG) of rickettsiae and are the causative agents of endemic and epidemic typhus, respectively. We recently generated a monoclonal antibody (BNI52) against R . typhi . Characterization of BNI52 revealed that it specifically recognizes TG rickettsiae but not the members of the spotted fever group (SFG) rickettsiae. We further show that BNI52 binds to protein fragments of ±30 kDa that are exposed on the bacterial surface and also present in the periplasmic space. These protein fragments apparently derive from the cytosolic GroEL protein of R . typhi and are also recognized by antibodies in the sera from patients and infected mice. Furthermore, BNI52 opsonizes the bacteria for the uptake by antigen presenting cells (APC), indicating a contribution of GroEL-specific antibodies to protective immunity. Finally, it is interesting that the GroEL protein belongs to 32 proteins that are differentially downregulated by R . typhi after passage through immunodeficient BALB/c CB17 SCID mice. This could be a hint that the rickettsia GroEL protein may have immunomodulatory properties as shown for the homologous protein from several other bacteria, too. Overall, the results of this study provide evidence that GroEL represents an immunodominant antigen of TG rickettsiae that is recognized by the humoral immune response against these pathogens and that may be interesting as a vaccine candidate. Apart from that, the BNI52 antibody represents a new tool for specific detection of TG rickettsiae in various diagnostic and experimental setups.
Publisher: American Thoracic Society
Date: 15-03-2019
Location: Australia
No related grants have been discovered for Marcel Kwiatkowski.