Homology based structure prediction (Phenylketonuria)

Summary

...

Model calculation

Lab journal
In our last task we already found some protein structures we now can discover a bit more. Therefore we first build a dataset with four structures with a sequence identity bigger than 60% and another also with four structures and a sequence similarity between 19 and 36% relative to our reference structure 2pah.

Modeller

...

Single template modeling

</figure> </figure> </figure>

Structure of target 2pah (green) and template 1j8u (purple) superimposed. For the creation of the superimposed structure Modeller was used.

Structure of target 2pah (green) and template 2phm(purple) superimposed. For the creation of the superimposed structure Modeller was used.

Structure of target 2pah (green) and template 3luy (purple) superimposed. For the creation of the superimposed structure Modeller was used.

<figure id="mod_1j8u">

<figure id="mod_2phm">

<figure id="mod_3luy">

2pah shows the 3D structure of the protein P00439 residues 118-452. Therefore only the biopterin domain is included not the ACT domain. It includes the catalytic domain. This is the part which gets aligned with the other structures in the figures above.
A quite good structure consensus was reached with 1j8u which was expected as it has a sequence identity of 100% to 2pah (<xr id="mod_1j8u"/>). It almost matches perfectly with the catalytic domain even forms the binding site. Only at the end the structure is not coiled any more and differs from 2pah.
2phm with a sequence identity of 89.7% also was aligned into the catalytic domain with quite good result although it is a bit worse than 1j8u (<xr id="mod_2phm"/>. The binding site is visible, but smaller and therfore more difficult to access.
For the structure 3luy, which only has a sequence identity of 22%, a lot of differences between the two structures can be viewed (<xr id="mod_3luy"/>). The binding site is completely missing, but again the structure is linked to the binding site of 2pah.

Multiple alignments

</figure> </figure> </figure>

Structure of target 2pah (green) and alignment of templates 1j8u, 2phm, 1qey and 1phz (purple) superimposed. For the creation of the superimposed structure Modeller was used.

Structure of target 2pah (green) and alignment of templates 1j8u, 2phm, 3luy and 1wyp (purple) superimposed. For the creation of the superimposed structure Modeller was used.

Structure of target 2pah (green) and alignment of templates 3luy, 1wyp, 2v27 and 2qmx (purple) superimposed. For the creation of the superimposed structure Modeller was used.

<figure id="mult_1j8u">

<figure id="mult_2phm">

<figure id="mult_3luy">

Swissmodel

SwissModel ... <ref name="swissmodel"> Konstantin Arnold, Lorenza Bordoli, Jürgen Kopp and Torsten Schwede (2006): "The SWISS-MODEL workspace: a web-based environment for protein structure homology modelling". Bioinformatics Vol.22(2): 195-201. doi:10.1093/bioinformatics/bti770 </ref>

</figure> </figure> </figure>

Structure of target 2pah (green) and template 1j8u (purple) superimposed. For the creation of the superimposed structure Swissmodell was used.

Structure of target 2pah (green) and template 2phn (purple) superimposed. For the creation of the superimposed structure Swissmodell was used.

Structure of target 2pah (green) and template 3luy (purple) superimposed. For the creation of the superimposed structure Swissmodell was used.

<figure id="swiss_1j8u">

<figure id="swiss_1j8u">

<figure id="swiss_3luy">

...

iTasser

iTasser... <ref name="itasser"> Yang Zhang (2008): "[zhanglab.ccmb.med.umich.edu/papers/2008_2.pdf I-TASSER server for protein 3D structure prediction]". BMC Bioinformatics Vol.9: 40. doi:10.1186/1471-2105-9-40 </ref>

We used different cutoff values for the modelling:

• for 1J8U we did not set any cutoff since the sequence has a identity of 100%
• for 2PHM we used a cutoff of 90% (all templates are excluded, which have a sequence identity >90% to the query protein)
• for 3LUY we used a cutoff of 35% (exclude all templates which have a sequence identity >35% to the query protein)

Model evaluation

Scores

//TODO How: Calculate the C_alpha RMSD of the models (use /mnt/project/pracstrucfunc13/bin/sap).??? <figtable id="eval_scores">

...
Template	Seq.Identity	GDT	Cα RMSD	...
Swiss-Model
1j8u	100	?	?
2phm	91.93?	?	?
3luy	16.79?	...	?
iTasser
1j8u	100		?
2phm	89?		?
3luy	22		?
Modeller - Single template modeling
1j8u	100	?	?
2phm	89?	?	?
3luy	22	?	?
Modeller - Multiple alignments
1j8u,2phm,1qey,1phz	>60	?	?
1j8u,2phm,3luy,1wyp	...	?	?
3luy,1wyp,2v27,2qmx	<30	?	?

</figtable> Swiss-Model:

• 1j8u:
  • Ligands in the template: FE2: 2, H4B: 2.
  • Ligands in the model: FE2: 2
• 2phm:
  • Ligands in the template: FE: 2.
  • Ligands in the model: FE: 2
• 3luy:
  • Ligands in the template: PPY: 1.
  • Ligands in the model: none.

References

<references/>