Session ii g2 overview protein modeling mmc
Download as PPTX, PDF3 likes756 views
This document discusses protein homology modeling. It begins with an introduction to proteins, amino acids, and protein structure. Homology modeling is introduced as a method to predict the 3D structure of a target protein based on its sequence and the known structure of a homologous template protein. The document then outlines the main steps in homology modeling: template selection using BLAST, target-template alignment, model construction using software like MODELLER, and model validation using tools like Errat. It also discusses commonly used protein structure prediction servers like SwissModel, I-TASSER, HHpred and provides an example workflow for selecting the best model.
Session ii g2 overview protein modeling mmc
- 2. Introduction • What is protein? - providing energy - building/repairing tissue in body • Amino Acids - protein is made up of amino acids. - 20 standard amino acids
- 3. Introduction • Sequence - string represent the primary structure of the protein Primary structure 3D structure
- 4. Introduction • Homology Modeling refers to constructing an model of the "target" protein from its amino acid sequence and/or an experimental 3D structure (the "template") ?
- 5. Introduction • Homology Modeling o When we want to know the 3D structure of a protein(target) that has not been solved o You have only the sequence o If there exist a solved experimental 3D structure(template) is similar to target protein o We can use software to arrange the backbone of the sequence.
- 6. Model Production Step 1 - Template Selection Step 2 - Target-Template Alignment Step 3 - Model Construction Step 4 - Model Validation
- 7. Step 1. Template Selection • First step in homology modeling is the identification of the best template structure. • The simplest method of template identification relies on database search techniques such as BLAST.
- 8. Step 1. Template Selection • BLAST(Basic Local Alignment Search Tool) Input: sequence Output: templates with Identities, E-value …
- 9. Step 2. Target-Template Alignment • Comparison between target & template. • Used to generate the model of protein.
- 10. Step 3. Model Construction • Fragment assembly relied on the assembly of a complete model from closely related solved structures. • Segment matching divides the target into short segments, matched to its own template fitted from the Protein Data Bank. • Spatial restraint based common software used is MODELLER.
- 11. Step 4. Model Validation • An example: Tool: Errat
- 12. Model Selection The Protein Model Portal • ModWeb • M4T • Swiss Model • I-TASSER • HHpred
- 15. Swiss Model
- 16. Model Selection Server Pdb files Model Name Hhpred 1 (Chun)HHpred I-TASSER 3 (Chun)I-TASSER model1 C-score=1.13 (Chun)I-TASSER model2 C-score=-1.39 (Chun)I-TASSER model3 C-score=0.70 M4T 1 (Chun)M4T ModelWeb 2 (Chun)ModelWeb Model_fd (Chun)ModelWeb Model_f5 Swiss Model 1 (Chun)SwissModel
- 18. Model Selection Model Name Errat(overall quality factor) Verify 3D( percentage of the residues had an averaged 3D-1D score > 0.2) Prove(Z score) (Chun)HHpred 86.856 93.52% 0.349 (Chun)I-TASSER model1 45.153 93.52% -0.075 (Chun)I-TASSER model2 48.329 77.31% 0.042 (Chun)I-TASSER model3 52.806 87.78% -0.176 (Chun)M4T 77.041 92.52% -0.137 (Chun)ModelWeb Model_fd 77.836 96.12% 0.050 (Chun)ModelWeb Model_f5 76.350 91.46% 1.196 (Chun)SwissModel 85.492 93.42% 0.302
- 19. Conclusion • Homology modeling o Input sequence -> Template -> Build Model -> Validation • Need understand and use tools well • It cost time to validate model
- 20. Questions?
Editor's Notes
- #3: First, I will explain some keywords in PHM and give you an introduction about what PHM is. Protein is an important and essential nutrient for everyone.All kinds of proteins in biological body is made up of….
- #8: In introduction we said that Homology Modeling build the model from amino acid sequence and experimental 3D structure. The experimental 3D structure we said there is the template.
- #9: Basically in this tool, your input is sequence, then the tool will search the database to find the best template. sometimes several templates are produced for a single query sequence, but we only choose the most likely candidate.
- #10: You can see it compares target and template sequence.This target template alignment are used to generate the 3D model of protein.
- #13: After we talk about model construction, Next part I am going to talk about is model chosen. PMP provides access to various server for model building, and quality assessment. I use PSI to get and collect pdb files for input sequence.
- #14: This figure is what protein model portal uses for structure modeling. We can see steps with blue stars are homology modeling steps what we are using.Besides it, the portal provides more like structure prediction when no template found. Or visualization of a model after built it etc.
- #16: After we submit the sequence and wait about one or two hours, the result will sent to you by email. Depends on different server, sometime they just send you pdb files, or sometimes they send you a link to a page contains model and some validation.Here we take Swiss Model as an example. In Model summary page we can see template selection stuff, like xxxx. And also some model quality parameter, xxxx.The model it produce we can download as many different format like, xxxx.
- #17: This table shows the result I get from protein model portal. I got 8 models from one input sequence from these 5 server
- #18: There are lots of server existing for model building. The input sequence using different server will creates many pdb files.We need to choose the best model from them, but HOW?
- #19: To decide which model is best we need test the pdb file on validation tools. There are lots of tools but here I just list three here. They are xxxx,All those parameters have different meaning. Basically, we need to understand what value is a good, what value is bad for each tool. To choose the best model, we need to test many validation tools and compare the quality parameters.