Finding a local homology domain by BLAST and pairwise alignment

a) Find the sequence for the human DNA repair gene Xrcc1 in SwissProt and create a local copy.

b) Perform a blast search against the non redundant protein database. **NOTE**: this sequence contains low complexity regions, therefore use blast's filter option, otherwise lots of junk will be reported

c) By using pairwise alignment, check which of the matches are significant and which are potentially interesting. Check the high scoring matches for putative biological relevance and note the results.

d) Find the conserved region of Xrcc1 and Rad4 by using dotplots. (Use Window=35, Stringency=20) Note the internal repeat in Rad4.

---

---

Multiple alignment of homology domains an profile searches

a) use pileup to create a multiple alignment of these three sequence fragments (in MSF format)

b) use lineup to look at the multiple alignment and to remove potential non-conserved overhanging ends. Save the edited multiple alignment in MSF format.

c) use profilemake to create a profile from the edited MSF file. Create the profile with lineas symbol weighting and stringent treatment of non-observed symbols.

d) use profilesearch to search the profile against a protein sequence database. Run this search with compositional averaging switched off. Don't forget to restrict the output list to the 100 best sequences, otherwise ALL SEQUENCES in the database will be reported ! Profile searches against big databases take a lot of time and cpu resources. The output file of a profilesearch against the complete SwissProt database will be provided.
NOTE: use the substitution matrix blossum45.cmp  with the option -data=genmoredata:blosum45.cmp)

e) Check the output list for high scoring matches of potential relevance. Create profile-to-sequence alignments with selected examples using profilesegments or profilegap.

Advanced exercises

The following exercises might be too time consuming to perform now -maybe try them at home!

f) Use profilegap to extract from the high-scoring sequences those fragments that match with the profile.

g) Add these new sequence segments to the existing multiple alignment, either by pileup or by manual editing.

h) Repeat the profilemake/profilesearch cycle iteratively as described above.

The Psi-Blast Server

a) Reproduce the results of the previous exercise using the NCBI Psi-Blast Server.

b) Collect as many members as you wish and produce a multiple alignment.

c) Produce a multiple alignment convincing nought for publication, using the available tools.

Here you go!if this domain was new, this would be a paper in TIBS :-)