|
|
Hidden Markov models as a tool for protein secondary structure prediction
Kraus, Ondřej ; Novotný, Marian (advisor) ; Mokrejš, Martin (referee)
Hidden Markov models are ideal tool for sequence analysis therefore they are used also for protein secondary structure prediction. A number of tools for protein secondary structure prediction exist today a part of them utilizes also hidden Markov models. Hence I try to introduce them to a reader and explain him the way they work and their advantages and disadvantages in this assay. The majority of methods predict three secondary structures with accuracy between 60% and 80% nevertheless with regard to different testing methodologies the results should be treated solely as indicative. Key-words: protein structure prediction, hidden Markov model, secondary structure
|
|
|
Modern methods for protein secondary structure prediction and their comparison
Kraus, Ondřej ; Novotný, Marian (advisor) ; Pleskot, Roman (referee)
Today, there are several protein secondary structure predictors; most of them use algorithms such as hidden Markov models or artificial neural networks. Therefore I will introduce them to a reader in my thesis. I will explain their principles, as well as their advantages and disadvantages. The majority of contemporary predictors have accuracy 70%-80% for prediction of three types of protein secondary structure. However these results are only approximate, due to different testing methodology. Therefore the user should get familiar with the method and its testing methodology in detail at first. Key-words: protein structure prediction, hidden Markov model, artificial neural network, nearest neighbour, protein secondary structure
|
|
|
Are there any sequence determinants of functional divergence of GTPases?
Kraus, Ondřej ; Novotný, Marian (advisor) ; Potocký, Martin (referee)
Small GTPases are important proteins that affect many cellular processes. In my work I compare the five most important protein families of small GTPases - Arf, Rab, Ran, Ras and Rho to identify amino acids responsible for major functional differences between different protein families. To compare them, I have used the structural data from the PDB database and sequences from the UniProt database. I have discovered previously undescribed groups of amino acids specific for each protein family of small GTPases with the help of programs ConSurf and Sca5. I also carried out a pilot study of the applicability of B-factors as indicators of bond strength in the protein structure on the example of small GTPases. The first results are not entirely conclusive, but they do not exclude the applicability of B-factors as indicators of bond strength either. Powered by TCPDF (www.tcpdf.org)
|
|
|
Modern methods for protein secondary structure prediction and their comparison
Kraus, Ondřej ; Novotný, Marian (advisor) ; Pleskot, Roman (referee)
Today, there are several protein secondary structure predictors; most of them use algorithms such as hidden Markov models or artificial neural networks. Therefore I will introduce them to a reader in my thesis. I will explain their principles, as well as their advantages and disadvantages. The majority of contemporary predictors have accuracy 70%-80% for prediction of three types of protein secondary structure. However these results are only approximate, due to different testing methodology. Therefore the user should get familiar with the method and its testing methodology in detail at first. Key-words: protein structure prediction, hidden Markov model, artificial neural network, nearest neighbour, protein secondary structure
|
|
|
Hidden Markov models as a tool for protein secondary structure prediction
Kraus, Ondřej ; Novotný, Marian (advisor) ; Mokrejš, Martin (referee)
Hidden Markov models are ideal tool for sequence analysis therefore they are used also for protein secondary structure prediction. A number of tools for protein secondary structure prediction exist today a part of them utilizes also hidden Markov models. Hence I try to introduce them to a reader and explain him the way they work and their advantages and disadvantages in this assay. The majority of methods predict three secondary structures with accuracy between 60% and 80% nevertheless with regard to different testing methodologies the results should be treated solely as indicative. Key-words: protein structure prediction, hidden Markov model, secondary structure
|
|
|
Analýza vybraného sortimentu spoločnosti IKEA z hľadiska hodnoty pre zákazníka
Medvecová, Gabriela ; Vlček, Radim (advisor) ; Kraus, Ondřej (referee)
Teoretická část se zaobírá pojmami jako hodnota pro zákazníka, komerčná úspěšnost ve vztahu k hodnotě pro zákazníka. Pojednává se taky o teorii funkce a funkčního přístupu. Jsou uvedeny různé metody stanovení hodnoty významu funkcí, metody stanovení stupně splnění funkcí a metody stanovení nákladů na tyto funkce. V praktické části se zjišťuje závislost mezi velikostí objemu prodejú vybaných matrací Sultan společnosti IKEA na ukazateli hodnoty pro zákazníka.
|
guest ::
RSS feed.