sequenenz_analyzer

Introduction

This project is university project. In this project we were provided with a data set. The data set contained several protein prediction-embeddings. In addition, we had for each protein the information if it the protein has an exophysiologial(exo.) function or an endophysiological(endo.) function.

After conducting a normal EDA, we were trying to analyse the data in a proper way. Therefore, I came up with the idea to analyse all embeddings which are to a prediction for each amino acids. To find then a combination or a specific amino acids which has a great evidence for each amino acid to be exo or endo. This idea tries to find a new way to look at each specific amino acids and the connected properties of the amino acid.

Wrangling with the embedddings

In addition, I only looked at the following embeddings:

• Amino Acid sequence
• Dssp3 Prediction for each amino acid
• Dssp8 Prediction for each amino acid
• Conservation score for each amino acid
• Membran_tmbed score for each amino acid
• Metall Binding per amino acid
• Nucleic Acid Binding per amino acid
• Small Molecules Binding per amino acid

AA:                         V   G   V   L   L   D   I   L   Q   R   ...
dssp3:                      E   L   L   L   H   H   H   H   E   L   ...
dssp8:                      H   H   H   T   T   C   E   E   E   E   ...
conserv_score:              0   3   3   3   4   0   5   0   5   5   ...
membran_tmbed:              S   S   S   o   o   o   o   o   o   o   ...
metal_binding:              M   M   -   M   -   -   -   -   -   -   ...
nucleic_binding:            N   -   -   -   N   -   -   -   -   -   ... 
small_molecules_binding:    -   -   S   -   -   S   -   S   -   S   ...
position:                   0   1   2   3   4   5   6   7   8   9   ...

In the next phase, I was constructing for each amino acid a word by combining all the the letters under each position. Each word describes the amino acid more precisely. With the example above I was constructing the following words for position 0, 1 and 2:

0:  VEH0SMN-    |   Translated: Valine,Sheet,alpha-helix,cons:8,Transmembrane-alpha-helix,MetallBinding,NucleicBindingSmallMoleculeBinding
1:  GLH3SM--    |   Tranalated:...
2:  VLH3S--S    |   Translated:...

So, I was creating for each of the proteins their set of amino-acid-word. Therefore, I was collecting in addition the positions where the words occure and how many the word occured per data set.

Interpreting the data parts