Difference between revisions of "Gaucher Disease: Task 05 - Lab Journal"

From Bioinformatikpedia
(Modeller)
(Calculation of models)
Line 44: Line 44:
   
 
As the PIDE with P04062 in the low-PIDE group we have selected in task 4 is too low and 10.74% PIDE with 2GEP_A was not enough for Swiss-Model to align the sequences (with BLAST or HHsearch), we looked again at the found hit list in [[Gaucher Disease: Task 02 - Alignments|task 2]] (HHblits, 2 iterations against Uniprot20 followed by one iteration against pdb_full with E-value cutoff 10E-10). The set of structures finally used is:
 
As the PIDE with P04062 in the low-PIDE group we have selected in task 4 is too low and 10.74% PIDE with 2GEP_A was not enough for Swiss-Model to align the sequences (with BLAST or HHsearch), we looked again at the found hit list in [[Gaucher Disease: Task 02 - Alignments|task 2]] (HHblits, 2 iterations against Uniprot20 followed by one iteration against pdb_full with E-value cutoff 10E-10). The set of structures finally used is:
 
   
 
{| border="1" cellpadding="5" cellspacing="0" align="center"
 
{| border="1" cellpadding="5" cellspacing="0" align="center"
Line 75: Line 74:
 
|-
 
|-
 
|}
 
|}
 
TODO: brief execution explanation?
 
   
 
===Modeller===
 
===Modeller===
We sued the command line executable to rum Modeller. The necessary scripts for the alignments (pairwise and MSAs) and the modelling (with single and multiple templates) were prepared using this [https://i12r-studfilesrv.informatik.tu-muenchen.de/wiki/index.php/Using_Modeller_for_TASK_4 tutorial].
+
We sued the command line executable to run Modeller. The necessary scripts for the alignments (pairwise and MSAs) and the modelling (with single and multiple templates) were prepared using this [https://i12r-studfilesrv.informatik.tu-muenchen.de/wiki/index.php/Using_Modeller_for_TASK_4 tutorial].
   
 
===Swiss-Model===
 
===Swiss-Model===
 
 
We executed Swiss-Model [http://swissmodel.expasy.org/workspace/index.php?func=modelling_simple1&userid=USERID&token=TOKEN online] using the 'Automatic Modelling Mode'. In the advanced options, the specific template was specified.
 
We executed Swiss-Model [http://swissmodel.expasy.org/workspace/index.php?func=modelling_simple1&userid=USERID&token=TOKEN online] using the 'Automatic Modelling Mode'. In the advanced options, the specific template was specified.
   
 
===iTasser===
 
===iTasser===
 
 
We also used iTasser from the [http://zhanglab.ccmb.med.umich.edu/I-TASSER/ web-server]. In "Option I" a specific template was specified.
 
We also used iTasser from the [http://zhanglab.ccmb.med.umich.edu/I-TASSER/ web-server]. In "Option I" a specific template was specified.
 
<!--"Option II" should have been used to exclude homologous templates for the low similarity template groups-->
 
<!--"Option II" should have been used to exclude homologous templates for the low similarity template groups-->
  +
  +
==Evaluation of models==
   
 
===RMSD and GDT-score calculation===
 
===RMSD and GDT-score calculation===

Revision as of 01:28, 3 September 2013

This page is still under construction

Calculation of models

Structures set

We assembled the following structures in task 4, now we divide them into two groups: at > 60% and at < 30% sequence identity to our template protein, P04062 (536 aa long). PIDE was calculated as follows:

  • first we aligned the two fasta sequences with ClustalW
  • then calculated the PIDE (pairwise sequence identity) using SIAS with default options

Selected structures for single-template modelling are written in bold.

Homologous structures to P04062
PDB ID PIDE with target (%) Length (aa)
High PIDE
2XWD_A 91.51 505
2NSX_A 92.53 497
2NT1_A 92.53 497
Low PIDE
2GEP_A 10.74 497
2F7K_A 8.95 327
2QGU 5.59 211
2ISB_A 5.59 192
2DJF_A 3.17 119
2DJF_B 2.98 164
2DJF_C 4.47 69

As the PIDE with P04062 in the low-PIDE group we have selected in task 4 is too low and 10.74% PIDE with 2GEP_A was not enough for Swiss-Model to align the sequences (with BLAST or HHsearch), we looked again at the found hit list in task 2 (HHblits, 2 iterations against Uniprot20 followed by one iteration against pdb_full with E-value cutoff 10E-10). The set of structures finally used is:

Homologous structures to P04062
PDB ID PIDE of alignment with target (%) Length (aa) Aligned columns (aa) Query coverage
High PIDE
3KE0_A 100 497 496 41-536
2XWD_A 100 505 497 40-536
2WKL_A 100 497 496 41-536
2NSX_A 100 497 496 41-536
Low PIDE
2WNW_A 29 447 440 75-534
1VFF_A 22 423 98 151-228
3II1_A 20 535 84 452-514

Modeller

We sued the command line executable to run Modeller. The necessary scripts for the alignments (pairwise and MSAs) and the modelling (with single and multiple templates) were prepared using this tutorial.

Swiss-Model

We executed Swiss-Model online using the 'Automatic Modelling Mode'. In the advanced options, the specific template was specified.

iTasser

We also used iTasser from the web-server. In "Option I" a specific template was specified.

Evaluation of models

RMSD and GDT-score calculation

We used the TM-score tool from Zhang lab to calculate the RMSD and GDT-score. We also looked at the TM score. Execution:

TMscore <model> <native>

C_alpha RMSD calculation

We visualized all created models with Pymol. For this we aligned each model with the reference structures, 1OGS_A and 2V3E_B, and calculated the RMSD between the corresponding C_alpha atoms like in task 4:

align <native> and resi 1-497 and name ca, <model> and resi <from>-<to> and name ca

where "native" is the reference structure (1OGS_A or 2V3E_B) and "model" is the name of the model, "from" is its first residue and "to" its last residue.