Qucosa - Publikationsserver der Universität Leipzig
Not a member yet
16020 research outputs found
Sort by
Cell-free expression of the Growth Hormone Secretagogue Receptor and subsequent biophysical characterization by solid-state NMR
No full textМикола Лисенко та національні школи європейського романтизму (Mykola Lysenko und die nationalen Schulen der europäischen Romantik. Eine Monographie), hrsg. von Igor Pyliatyuk, Lviv: Halych Press, 2024. 432 S., Abb.
No full text
Entwicklung automatisierter Bestrahlungsprozesse zur Behandlung von Pathogenen
No full textIonisierende Strahlung wird auf vielfältige Weise in Bereichen der Industrie, Landwirtschaft, Medizin und in der Forschung eingesetzt, z.B. zur Sterilisation von Oberflächen und Medizinprodukten, Polymermodifikation, Diagnostik oder in der Entwicklung von Impfstoffen. Im Rahmen dieser Arbeit wurden Prozesse zur automatisierten Bestrahlung von flüssigen Pathogen-Proben mit niederenergetischer Elektronenstrahlung (eng. Low-Energy Electron Irradiation, LEEI) etabliert. LEEI, wie auch andere Arten der ionisierenden Strahlung, erzielt Effekte in biologischen Proben vorrangig durch Schädigung von Nukleinsäuren. Proteinstrukturen, unter anderem essentielle Antigene für eine effektive Immunantwort, bleiben größtenteils intakt. Kernstück der Arbeit ist eine Prototyp-Anlage, die sich am Fraunhofer-Institut für Zelltherapie und Immunologie befindet, die erstmalig die Bestrahlung von Flüssigkeiten in einem automatisierten, skalierbaren und produktionsgeeigneten Prozess ermöglicht. In diese Anlage können verschiedene Module integriert werden, die dünne Flüssigkeitsfilme generieren, da der Einsatz von LEEI stark durch die niedrige Eindringtiefe (<200 µm) limitiert ist. Fokus der Arbeit liegt auf der Entwicklung von Prozessen zur Inaktivierung vom Früh-Sommer-Meningoenzephalitis-Virus (FSME bzw. engl. TBEV) und dem humanen respiratorischen Synzytial-Virus (RSV), sowie zur Attenuierung (Reduktion der Virulenz eines Erregers unter erhalt der Immunogenität) zweier Endoparasiten der Gruppe Apicomplexa, Toxoplasma gondii und Cryptosporidium parvum
A study on the material properties of novel PEGDA/gelatin hybrid hydrogels polymerized by electron beam irradiation
No full textGelatin-based hydrogels are highly desirable biomaterials for use in wound dressing, drug delivery, and extracellular matrix components due to their biocompatibility and biodegradability. However, insufficient and uncontrollable mechanical properties and degradation are the major obstacles to their application in medical materials. Herein, we present a simple but efficient strategy for a novel hydrogel by incorporating the synthetic hydrogel monomer polyethylene glycol diacrylate (PEGDA, offering high mechanical stability) into a biological hydrogel compound (gelatin) to provide stable mechanical properties and biocompatibility at the resulting hybrid hydrogel. In the present work, PEGDA/gelatin hybrid hydrogels were prepared by electron irradiation as a reagent-free crosslinking technology and without using chemical crosslinkers, which carry the risk of releasing toxic byproducts into the material. The viscoelasticity, swelling behavior, thermal stability, and molecular structure of synthesized hybrid hydrogels of different compound ratios and irradiation doses were investigated. Compared with the pure gelatin hydrogel, 21/9 wt./wt. % PEGDA/gelatin hydrogels at 6 kGy exhibited approximately up to 1078% higher storage modulus than a pure gelatin hydrogel, and furthermore, it turned out that the mechanical stability increased with increasing irradiation dose. The chemical structure of the hybrid hydrogels was analyzed by Fourier-transform infrared (FTIR) spectroscopy, and it was confirmed that both compounds, PEGDA and gelatin, were equally present. Scanning electron microscopy images of the samples showed fracture patterns that confirmed the findings of viscoelasticity increasing with gelatin concentration. Infrared microspectroscopy images showed that gelatin and PEGDA polymer fractions were homogeneously mixed and a uniform hybrid material was obtained after electron beam synthesis. In short, this study demonstrates that both the presence of PEGDA improved the material properties of PEGDA/gelatin hybrid hydrogels and the resulting properties are fine-tuned by varying the irradiation dose and PEGDA/gelatin concentration
Milia en Plaque on the Right Ear-Lobe
No full textA milia en plaque on the right earlobe is seen, which was moved by excision under local anesthesia
Diaphanous-related formin subfamily: Novel prognostic biomarkers and tumor microenvironment regulators for pancreatic adenocarcinoma
No full textThe diaphanous-related formin subfamily includes diaphanous homolog 1
(DIAPH1), DIAPH2, and DIAPH3. DIAPHs play a role in the regulation of actin
nucleation and polymerization and in microtubule stability. DIAPH3 also
regulates the assembly and bipolarity of mitotic spindles. Accumulating
evidence has shown that DIAPHs are anomalously regulated during
malignancy. In this study, we reviewed The Cancer Genome Atlas database
and found that DIAPHs are abundantly expressed in pancreatic
adenocarcinoma (PAAD). Furthermore, we analyzed the gene alteration
profiles, protein expression, prognosis, and immune reactivity of DIAPHs in
PAAD using data from several well-established databases. In addition, we
conducted gene set enrichment analysis to investigate the potential
mechanisms underlying the roles of DIAPHs in the carcinogenesis of PAAD.
Finally, we performed the experimental validation of DIAPHs expression in
several pancreatic cancer cell lines and tissues of patients. This study
demonstrated significant correlations between DIAPHs expression and
clinical prognosis, oncogenic signature gene sets, T helper 2 cell infiltration,
plasmacytoid dendritic cell infiltration, myeloid-derived suppressor cell
infiltration, ImmunoScore, and immune checkpoints in PAAD. These data
may provide important information regarding the role and mechanisms of
DIAPHs in tumorigenesis and PAAD immunotherapy
Haplotype-based Methods for aDNA Data Analysis
No full textThis thesis focuses on new haplotype-based methods for analyzing ancient DNA
(aDNA) data. Methods based on allele frequency have so far dominated the
aDNA world and have been the primary workhorse that drives most of the
exciting discoveries from aDNA in the past decade. However, it suffers from
two shortcomings: first, it has low resolution for decoding the demography of
the recent past because the widely used in-solution capture enrichment panel
(1240k panel) mostly consists of common variants, which is not particularly
informative about recent population history; second, it can produce biased
results due to deep population structure.
Haplotype-based methods (e. g., imputation, phasing) and their downstream
tasks (e. g., identity-by-descent segments calling) have long been applied to DNA
data collected from modern populations and have greatly advanced many areas
of human genetics, including both medical and population genetics. However,
progress has been lagging in the aDNA world. This thesis presents three new
haplotype-based methods that open new avenues for aDNA data analysis.
Chapter 2 introduces hapCon, a new method for estimating contamination
rate in aDNA sequencing data that implicitly performs imputation on the haploid
male X chromosome. Because hapCon’s ability to use linkage disequilibrium
(LD) to draw information from neighboring sites and to use the Li&Stephen’s
haplotype copying model to draw information from a large reference panel, it
consistently outperforms previously published methods by producing estimates
with much lower variance. Section 2.4 also introduces how hapCon can be
extended to estimate contamination rate from runs of homozygosity (ROH),
which allows it to be applied to female samples that contain 50cM or more
ROH blocks as well.
Chapter 3 introduces ancIBD, a new method for detecting identity-bydescent (IBD) segments from aDNA data with 0.25x coverage (WGS data) or
1x coverage (1240k data). The model was designed by Harald Ringbauer, my
PhD supervisor, during his postdoc. I was mainly responsible for simulations
and extending it to detecting IBD2 segments and to IBD detection on X
chromosomes.
Chapter 4 presents TTNe, a new method to estimate effective population
size (Ne) trajectory from time-series IBD segments. This is particularly handy
for the aDNA community because samples collected from the same site often
date to different time periods. However, as most models are designed to work
with modern DNA, they assume all samples are contemporaneous. Using
simulations, we show that utilizing IBD sharing in time series has increased
resolution to infer recent fluctuations in effective population sizes compared
to methods that only use contemporaneous samples. Finally, we developed
an approach for estimating and modeling IBD detection errors in empirical IBD analysis. To showcase the practical utility of TTNe, we applied it to
two time transects of ancient genomes, individuals associated with the Copper
Age Corded Ware Culture (CWC) and Medieval England. In both cases, we
found evidence of a growing population, a signal consistent with archaeological
records.
Chapter 5 discusses possible directions for improving the works presented in
this thesis
Learning queries under description logic ontologies
No full textIn knowledge bases, relational data is combined with knowledge formalized
through logic. This formalized knowledge is often called an ontology. Queries
to knowledge bases are then answered with respect to the ontology, yielding
more comprehensive results.
Real-world knowledge bases contain thousands of relations and many thousand
logical statements in their ontology. Writing queries that yield the desired
answers therefore requires effort and expertise.
In this dissertation, we investigate the learnability of conjunctive queries
(CQs) under description logic (DL) ontologies. We focus on learning in the
sense of Angluin’s exact learning and of probably approximately correct (PAC)
learning, and on description logics of the EL and DL-Lite-core families, which
underlie the OWL 2 EL and QL profiles, respectively.
In both models of learning, the learner tries to learn a target query based on
limited available information. In exact learning, this information is provided
by a teacher who answers certain types of questions (membership queries and
equivalence queries) truthfully, whereas in PAC learning the learner receives
randomly drawn labeled data examples.
One key question is whether polynomial-time algorithms exist that allow the
learner to always learn the target query, even when the information about the
target query is provided with respect to a DL ontology. In this dissertation,
we aim to determine for which classes of CQs and for which ontology languages
such polynomial-time learning algorithms exist, and which kinds of questions
(membership queries and equivalence queries) are necessary for polynomial-time
learning in the exact learning model. For this, we build upon existing results
on the learnability of queries without ontologies. The following are the main
contributions of my thesis:
We show that membership queries alone suffice to learn unary acyclic connected
CQs (that correspond to concepts of the description logic ELI) in polynomial
time under DL-Lite-core ontologies. This result also holds if the ontology
includes role hierarchies and a limited form of functionality assertions.
In contrast, We show that EL ontologies (and most extensions of DL-Lite-core)
make learning with only membership queries difficult: in the worst case, an
exponential number of membership queries is required to identify the target
query.
We show that using both membership queries and equivalence queries, the class
of tree-shaped CQs (as well as the larger class of chordal and symmetry-free
CQs) is polynomial-time learnable under EL ontologies. This also holds for
unary acyclic connected CQs under DL-Lite-horn ontologies, an extension of
DL-Lite-core.
However, these results do not extend to ontologies formulated in ELI, which
extends EL with inverses. Already simple query classes are not polynomial-time
learnable under ELI ontologies.
We review that equivalence queries alone are not sufficient to learn simple
path-shaped CQs in polynomial time, unless NP= RP. This implies that no
polynomial-time PAC learning algorithms exist for CQs.
Instead, We show that PAC learning of CQs under ontologies, with a required
sample size that grows only polynomially, is possible using an algorithm that
always returns the smallest query that fits the labeled examples. We
implemented such an algorithm for tree-shaped CQs (which correspond to ℰℒ
concepts) under EL ontologies and show that the implementation compares
favorably to an existing EL concept learning algorithm on benchmarks