Lab Journal of Task 4 (MSUD)

From Bioinformatikpedia
Revision as of 16:29, 30 May 2013 by Schillerl (talk | contribs) (Use structural alignments to evaluate sequence alignments)

Explore structural alignments

Use structural alignments to evaluate sequence alignments

Creation of structure models from homologous sequences

The input for the the script hhmakemodel.pl was create with HHsearch:

hhsearch -i /mnt/home/student/weish/master-practical-2013/task01/refseq_BCKDHA_protein.fasta -d /mnt/project/pracstrucfunc13/data/hhblits/pdb70_current_hhm_db -o hhsearch_BCKDHA.hhr -Z 200000 -B 200000

The models were made as follows:

/usr/share/hhsuite/scripts/hhmakemodel.pl hhsearch_BCKDHA.hhr -d /mnt/project/pracstrucfunc13/data/pdb/20120401/entries/* -m 2 4 5 6 10 20 28 31 39 43 -ts hhmakemodel_BCKDHA.pdb

As can be seen in the option -m, 10 structures were chosen to make a model of each. The structures were selected preferably to cover the whole range of alignment scores (E-value, probability, sequence identity), but no hits with very high E-values were taken.

The models are stored in /mnt/home/student/schillerl/MasterPractical/task4/hhmakemodel/.

Comparison with known structure

The models were structually aligned with the reference structure 2BDF using LGA, which was run on this server with default parameters.

LGA results can be found in /mnt/home/student/schillerl/MasterPractical/task4/LGA/.

Stastitical evaluation of alignment scores (from sequence and structural alignments), especially calculation of correlation between these, was performed with R. See /mnt/home/student/schillerl/MasterPractical/task4/alignment_statistics.txt for an overview of the scores.