I am working on predicting the 3D structure of a protein with an amino acid sequence exceeding 3000 residues. Unfortunately, the full structure for this protein is not available in the Protein Data Bank (PDB). I have tried utilizing various bioinformatics tools such as Swiss-Model, PHYRE2, and I-TASSER, but they were unable to provide a complete structure for my protein due to its length.
I encountered a limitation with I-TASSER, as it only allows predictions for proteins with amino acid sequences up to 1500 residues, which falls short of my protein's length.
Could anyone recommend alternative methods, software, or approaches that are capable of handling longer protein sequences and can predict the 3D structure effectively? Your insights and suggestions would be greatly appreciated
Hey, Harris Jan,
Why can you try homology modelling by using the Modeller 10.4 software? It's very simple. Just blast your FASTA sequence in the blastp suite and find a suitable template using the query cover and sequence identity parameters. But before you check in on some things in Blastp webpage itself, select Standard > Database > Protein Data Bank (PDB), so you only focus on already-solved similar protein structures. Now, the similar protein structure you found through Blast is your template, and the sequence you have is your target. Using the Modeller 10.4 software, you can now predict your own 3-dimensional protein.
If you still find it harder to work on the Modeller tool, watch this: https://youtu.be/k_sa9pCzstU?si=kO_2OSodl3F2aqrl.
Cheers,
Mohamed Asif
You may try to split the protein into several domain and predict each domain using AlpahFold2 etc structure prediction software. If your computational resourece is enough ,you can directly predict the protein with all length
- Badges
- Science topic