RMSD is usually taken as below 2 but if the length of Proteins is too high or large, is it affect RMSD or not? if, then how much it vary from sequence to sequence
Hi:
When comparing 3D structures, one uses the root mean square deviation (rmsd) as a quantitative measure of similarity. What is important to keep in mind is that the statistic is calculated pairwise for a selected set of atoms and for a given rotation/translation. Software is readily available to help with this and can quickly identify the rotation/translation to give a minimum rmsd. Common atom choices are: backbone heavy atoms (N, CA, C) or all heavy atoms of the entire sequence or selected segments. As you mention, values below 2 are structurally similar. Two proteins of identical structure but randomly oriented (superimposing centers of mass) can have values of 3 Å or above.
If amino acid sequences are different, one should not compare atoms that are not in homologous positions in both structures. One can compare the parts of the side chains that are homologous or simply restrict the atom selection to the backbone. For example if Alanine is replaced by Serine in homologous positions, N, CA, C, O, CB can be compared. It can happen that 2 proteins have structurally homologous domains. If this is not taken into consideration when selecting the atoms/residues for comparison, similarity can go undetected. For a given rmsd, the longer the segments or the greater the fraction of residues contained in the comparison, the more structurally similar the two proteins can be considered. As contained in the name, the statistic is a "mean" value so the chosen sequence length and number of atoms is compensated for.
Hope this helps (got to be long-winded...). Good luck.
Hi:
When comparing 3D structures, one uses the root mean square deviation (rmsd) as a quantitative measure of similarity. What is important to keep in mind is that the statistic is calculated pairwise for a selected set of atoms and for a given rotation/translation. Software is readily available to help with this and can quickly identify the rotation/translation to give a minimum rmsd. Common atom choices are: backbone heavy atoms (N, CA, C) or all heavy atoms of the entire sequence or selected segments. As you mention, values below 2 are structurally similar. Two proteins of identical structure but randomly oriented (superimposing centers of mass) can have values of 3 Å or above.
If amino acid sequences are different, one should not compare atoms that are not in homologous positions in both structures. One can compare the parts of the side chains that are homologous or simply restrict the atom selection to the backbone. For example if Alanine is replaced by Serine in homologous positions, N, CA, C, O, CB can be compared. It can happen that 2 proteins have structurally homologous domains. If this is not taken into consideration when selecting the atoms/residues for comparison, similarity can go undetected. For a given rmsd, the longer the segments or the greater the fraction of residues contained in the comparison, the more structurally similar the two proteins can be considered. As contained in the name, the statistic is a "mean" value so the chosen sequence length and number of atoms is compensated for.
Hope this helps (got to be long-winded...). Good luck.
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Proteins