|
|
Correction of the concept of drift in prediction models
Michálková, Eva ; Provazník, Valentine (referee) ; Schwarzerová, Jana (advisor)
The main goal of this bachelor thesis is the analysis of concept drift in metabolomics. Concept drift is an undesirable phenomenon and can be caused by nonstationary data. It can have a negative impact on the performance and reliability of predictive modelling. This challenge can be solved by concept drift detection and subsequent correction. One of the fields where this issue has recently emerged is metabolomic diagnostics. Metabolomic data analysis can lead to early detection of several serious diseases, which can help with the recovery process. When diagnosing an illnes predictive models present a way to make the process more efficient, faster and give the option of personalization. The first part of this thesis specifies concept drift, it’s detection and correction methods and the importance of metabolomics and prediction models. The second part deals with the implementation of some available algorithms for concept drift detection and correction and the implementation of automatic concept drift correction. Finally, in the second part results and their discussion are described.
|
|
|
Analysis of horizontal transfer of genetic components using static network analysis
Labanava, Anastasiya ; Jurečková, Kateřina (referee) ; Schwarzerová, Jana (advisor)
The bachelor's thesis focuses on the issue of horizontal genetic elements transfer between bacteria of different strains and the software analysis implementation that enables horizontally transferred genes identification. The packages and tools used were tested on a dataset of bacterial genomes from several strains. The thesis’ theoretical part provides a detailed description of the genetic components transfer between bacteria and describes modern laboratory techniques that enable genome sequencing in various ways. In the practical part, the thesis deals with the preprocessing of genomic files to obtain suitable data for annotation. To detect the horizontal transfer of genetic elements between bacteria, a script is introduced, which organizes annotated bacteria to tables and searches for the same genes in their genomes that, under theoretical assumptions, were horizontally transferred. Furthermore, the gene transfer is visualized using tools that graphically represent phylogenetic relations between bacteria. In the final step, bacterial genomes are connected into networks, and based on their static analysis, a discussion is conducted on the results accuracy and the success of the proposed analysis.
|
|
|
RNA secondary structure prediction
Polzerová, Nikola ; Schwarzerová, Jana (referee) ; Jurečková, Kateřina (advisor)
Sekundární struktura RNA se stala v posledních letech fenoménem. Hraje klíčovou roli v pochopení principů genové exprese a stability RNA, a také pokládá důležitý základ pro správnou predikci terciálních struktur. Proto dochází k velice rychlému rozvoji predikce pomocí výpočetních metod a také pomocí metod strojového učení. V rámci práce byly popsány nejčastěji citované metody strojového učení pro predikci sekundárních struktur RNA. Na základě získaných znalostí byla implementována reziduální neuronová síť. Implementovaná reziduální síť byla netrénována, validována a otestována na sekvencích z datasetu bpRNA-1m, jejichž sekundární struktury neobsahují pseudouzly.
|
|
|
Operon identifier: Identification of operon structures in the whole genome
Nejezchlebová, Julie ; Schwarzerová, Jana
Currently, operon prediction is based on the distance of neighboring genes on the functional relationships of their products that encode proteins in a given nucleotide sequence, or on ORF distances. This study deals with the design of a new function that detects operon structures based on information from gene expression or alternatively in combination with previous information from current online available tools. The function was implemented in Python language and tested on Clostridium beijerinckii NRRL B-598. This bacterium has huge potential in biotechnology and research due to its fermentation product, butanol.
|
|
|
Operon structures inference in genome-wide analysis
Nejezchlebová, Julie ; Jurečková, Kateřina (referee) ; Schwarzerová, Jana (advisor)
The bachelor thesis is devoted to the problem of derivation of operon structures and creation of a software tool that allows prediction of operon structures. The tool both predicts operons based on gene expression information, but also refines already predicted operons with gene expression information. The tool is tested on the bacteria Escherichia coli BW25113 and Clostridium beijerinckii NRRL B-598. The theoretical part is devoted to description of operon structure and function, genome sequencing, transcriptome analysis, Clostridium beijerinckii NRRL B-598, Escherichia coli BW25113 and already available online tools for inferring operon structures. In the practical part of the thesis, the pre-processing of raw transcriptomic data to obtain a suitable format for the prediction of operon structures, testing of online tools and the actual implementation of the tool itself are discussed.
|
|
|
Concept drift in metabolomic analysis
Koštoval, Aleš ; Provazník, Ivo (referee) ; Schwarzerová, Jana (advisor)
This bachelor thesis deals with machine learning, specifically the analysis of the concept drift. This is an unwanted phenomenon that can be detected in predictive models. Through detection followed by correction of the concept drift, predictive models become more reliable and can respond adequately to input data representing dynamic information. Metabolomic data can be considered a suitable representative of such data. Metabolomic data and their analysis can help to detect diseases such as diabetes mellitus or cancer early. In the first part of this bachelor thesis, the theoretical background of concept drift analysis and metabolomics analysis are described. The second part discusses the process of modeling predictive classifiers and implementing algorithms for concept drift detection. The practical part of the work was implemented in the Python programming language. Finally, the second part describes the results obtained and their discussion.
|
|
|
Dynamic metabolomic prediction from genetic variation
Nemčeková, Petra ; Weckwerth, Wolfram (referee) ; Schwarzerová, Jana (advisor)
Hordeum vulgare, tak ako mnoho ďalších plodín, trpí redukovaním genetickej rôznorodosti spôsobeným klimatickými zmenami. Preto je potrebné zlepšiť účinnosť jeho kríženia. Oblasť záujmu sa v poslednej dobe obracia na výskum nepriamych selekčných metód založených na výpočetných predikčných modeloch. Táto práca sa zaoberá dynamickou metabolomickou predikciou založenou na genomických dátach, ktoré pozostávajú z 33,005 jednonukleotidových polymorfizmov. Metabolomické dáta zahŕňajú 128 metabolitov 25 rodín Halle exotického jačmeňa. Hlavným cieľom tejto práce je vytvoriť metabolomické predikcie dynamických dát pomocou rôznych metód, ktoré boli vybrané na základe rôznych publikácií. Vytvorené modely napomôžu predikcii fenotypu alebo odhaleniu dôležitých vlastností rastliny Hordeum vulgare.
|
|
|
Metabolite Genome-Wide Association Studies Of Arabidopsis Thaliana
Schwarzerová, Jana
Current research based on the edge of bioinformatics and ecology engineering has hugepotential due to combination of laboratory analyses and advanced bioinformatics algorithms.The paper deals with a combination of GC-MS and LC-MS based metabolomic analysis for identificationand quantitation of metabolites in environmental perturbations with advanced bioinformaticsapproach of metabolite genome-wide association studies (mGWAS). This complex view is appliedto Arabidopsis thaliana. The main goal is to obtain genetic predictions focused on A. thalianaunder different environmental conditions. Currently, important ecological issues such as climatechange, pollution etc. have impact on the change of environment. It has a great effect on plantswhich serves as producers of oxygen or food. While simple observation reveals only a phenotypechange, changes in genotypes of organisms can be captured using mGWAS and further utilized inindustrial ecology and biotechnology.
|
|
|
Operon-Expresser: The Innovated Gene Expression-Based Algorithm For Operon Structures In-Ference
Schwarzerová, Jana
Current biotechnological research of bacterial genomes has huge potential due to the useof next-generation sequencing (NGS) platforms. NGS era opens paths for analysis of data for sufficientdescription of microorganisms with ecology and biotechnology potential in the future. Althoughsome tools for inference of specific structures in bacterial genomes exist, their pipelines andmethodologies are often based only on searching bacterial genome databases and comparison withmodel microorganisms. This paper deals with the design of a new algorithm for operon structuresinference in bacterial genomes. The algorithm combines searching in bacterial databases and geneexpression information processing. The algorithm was implemented in R language and tested onClostridium beijerinckii NRRL B-598. The bacterium is a typical performer in the field of biofuelsproduction due to its ability to produce butanol. Thanks to that this bacterial organism can be ofgreat potential from an ecological and biotechnological point of view. The paper also provides acomparison of operon structures derived by Operon-mapper and its extending by Operon-expresser.
|
|
|
Reproducible Analytical Pipeline For Using Raw Rna-Seq Data From Non-Model Organisms
Schwarzerová, Jana
Current biotechnological research of bacterial or archaeal genomes has huge potential due to the use of the next generation sequencing (NGS) platforms. NGS era unravelled huge analysis data for sufficiency description microorganisms with ecology potential in future. Nowadays, efforts lie in creating comprehensive pipelines that can be used for pre-processing analysis to enable effective following steps of high throughput data processing. This paper deals with design of data analysis pipeline for using raw RNA-Seq data that was applied to the Clostridium beijerinckii NRRL B-598. The bacterium is typical performer in the field of biofuels production thanks to its ability to produce butanol. Unfortunately, it is non-model organism as many other microorganisms which can be of great potential from ecological point of view. The proposed pipeline offers to take necessary steps in initial data processing that produces data of comparable quality to widely studied model organisms. Therefore, it can be combined with following pipelines for gene regulatory network inference, which was up to date matter of non-model organisms.
|
guest ::
