Dog kinesin

Canis familiaris (dog)

-e!Ensemble Dog BlastView was employed at ensembl.org to align the sequence of D. melanogastor kinesin heavy chain (P17210) against the Canis familiaris genome.The paramaters for the search were as follows: E =<0.5 filter="seg," length =" 3," hitdist =" 40," extension =" 2:1," threshold =" 15,">

-Results: 801 alignments, S values: 79-2828, E values: 0-0.0036

- e!Ensemble having returned too many redundant or inconsequential alignments, a blastp search of the NCBI database was performed with a query of aa's 1-350 of D. melanogaster KHC (P17210). The blastp algorithm proved to be much more selective, returning alignments of only 10 dog kinesin isoforms all with values of 1e-135 or better: Taxonomy Report. However, it was expected that this search was too selective, and that the use of a position-specific scoring matrix would better achieve a balance between selectivity and sensitivity and correctly identify the known dog proteins homologous with fruit-fly kinesin.

- A psi-BLAST search of aa's 1-350 of P17210 against the dog genome, with max E set at 0.005, returned 88 hits within the set paramaters (max E = 1e-4) and 2 outside of it (E = 3.6 and 6.3)...for some reason hitting "2nd iteration" does nothing but it shouldn't matter since the E values of hits 88 and 89 are so far apart. Also, I couldn't paste a working link to the psi-BLAST results into the blog, so I hope you didn't have your hopes too high for that.

*Overall, psi-BLAST proved to be the most useful tool for running an organism-specific search for a specific protein (dog, kinesin). Its usefulness was deemed to be its achievement of a balance avoiding excessive sensitivity (Ensemble) and excessive selectivity (blastp).