Dear Deepa, When classifying structures you need to first ask "why?". For example, if you want to get an idea about evolutionary relations, you must try to find their DNA sequences and analyse those in the light of a multiple sequence alignment (MSA) that follows the protein alignment. In pharma you might want to superpose them based on just the residues around the pocket of interest. If interested in regulatory processes, you map the regulation-related sequence motifs on a protein sequence MSA (that might be supported by a multiple structure superposition).
Your question contains insufficient detail for better answers. One question, though: What is wrong with DALI giving you 50-60 groups? You can combine groups in bigger groups, cann't you? If something crashes on 300+ structures then make a phylogenetic tree first (based on protein sequences) and remove doubles or near-doubles till you have less then 300 structures. Proteins that are >95% sequence identical normally are 'the same anyway'. Gert -------- Forwarded Message -------- Subject: [ccp4bb] Structural classification Date: Fri, 24 Apr 2015 13:36:42 +0100 From: Deepa Raju <deepakmraj...@gmail.com><mailto:deepakmraj...@gmail.com> Reply-To: Deepa Raju <deepakmraj...@gmail.com><mailto:deepakmraj...@gmail.com> To: CCP4BB@JISCMAIL.AC.UK<mailto:CCP4BB@JISCMAIL.AC.UK> Dear all, I am Deepa, MSc final year student. For my final year project, I am working on protein structure classification. I have a set of 300 methyltransferase protein structures. I wish to classify the proteins in to different groups based on their secondary structures and sequence. I have tried to classify the proteins based on sequence but it is not satisfactory in view point of structures. I have used the DALI server results like Z-score, RMS deviation for grouping. I ended up with 50-60 groups. I have used the PIRSF, but in that most of the structures are not available. I have also used Chimera - structure based sequence alignment, but the program is crashing for 300 structures. Any suggestions will be highly beneficial to me. Thank you in advance. With regards Deepa Het Radboudumc staat geregistreerd bij de Kamer van Koophandel in het handelsregister onder nummer 41055629. The Radboud university medical center is listed in the Commercial Register of the Chamber of Commerce under file number 41055629.
