|
|
Informační technologie v psychologii
Ižák, Tomáš ; Fapšo, Michal (oponent) ; Smrž, Pavel (vedoucí práce)
Tato práce se zajímá o možnosti zjištění charakterových vlastností autora textu za použití výpočetní techniky. Práce je založena na metodě manželů Kreitlerových, proto se zabývá zejména rozdělením vět na významové jednotky, přiřazení dimenze významové jednotce a získání charakterových rysů autora (neuroticismu a extroverze) z dimenzí významových jednotek.
|
|
|
Detection of Correlated Mutations
Ižák, Tomáš ; Bendl, Jaroslav (oponent) ; Martínek, Tomáš (vedoucí práce)
This work explores existing possibilities and methods of correlated mutations detection in proteins. At the beginning a theoretical background into explored area is provided. Exploitation of detected correlated mutations lies in a protein's tertiary structure prediction or searching functionally important sites. A state-of-the-art of existing tools and methods follows. In this work, methods based on statistics (for example Pearson correlation coefficient or Pearson's chi^2 test), Information theory (Mutual information - MI) and likelihood models (ELSC or Spidermonkey) are examined. The next part is devoted to the searching for an optimal algorithm for correlated mutations detection. To combine results from multiple different algorithms, is proposed as an optimal solution. It is also advised to exploit physico-chemical properties of amino acids during the detection. In practical part, the algorithm for detection of correlated mutations was developed. It is based on physico-chemical properties of amino acids and phylogenetic trees. Results gained using this method were compared with results gained from CAPS, CRASP and CMAT tools.
|
|
|
Informační technologie v psychologii
Ižák, Tomáš ; Fapšo, Michal (oponent) ; Smrž, Pavel (vedoucí práce)
Tato práce se zajímá o možnosti zjištění charakterových vlastností autora textu za použití výpočetní techniky. Práce je založena na metodě manželů Kreitlerových, proto se zabývá zejména rozdělením vět na významové jednotky, přiřazení dimenze významové jednotce a získání charakterových rysů autora (neuroticismu a extroverze) z dimenzí významových jednotek.
|
|
|
Detection of Correlated Mutations
Ižák, Tomáš ; Bendl, Jaroslav (oponent) ; Martínek, Tomáš (vedoucí práce)
This work explores existing possibilities and methods of correlated mutations detection in proteins. At the beginning a theoretical background into explored area is provided. Exploitation of detected correlated mutations lies in a protein's tertiary structure prediction or searching functionally important sites. A state-of-the-art of existing tools and methods follows. In this work, methods based on statistics (for example Pearson correlation coefficient or Pearson's chi^2 test), Information theory (Mutual information - MI) and likelihood models (ELSC or Spidermonkey) are examined. The next part is devoted to the searching for an optimal algorithm for correlated mutations detection. To combine results from multiple different algorithms, is proposed as an optimal solution. It is also advised to exploit physico-chemical properties of amino acids during the detection. In practical part, the algorithm for detection of correlated mutations was developed. It is based on physico-chemical properties of amino acids and phylogenetic trees. Results gained using this method were compared with results gained from CAPS, CRASP and CMAT tools.
|
host ::
RSS.