Using BLAST to Compare Nucleic Acid Sequences Starting link:

Exploring Genomes: Using BLAST to Compare Nucleic Acid Sequences
By Paul G. Young
Using BLAST to Compare Nucleic Acid Sequences
Starting link:
In the previous tutorial, “Learning to use BLAST”, you learned how to use BLAST to
search the genomic databases and to compare protein sequences. Proteins have a high
complexity and information content since there are 20 different possible amino acids at
each position in the sequence. As we have seen, there are also groups of amino acids with
related properties that are scored as being similar. Nucleic acids, on the other hand, have
only four possible choices at each position (AGCT[U]). For the most part, we only look
to match identical residues at a particular position in a DNA or RNA molecule. In this
tutorial we will start using BLASTN to compare the sequences of different transfer RNA
From the NCBI home page, choose ‘BLAST’ from the top menu, and then choose
‘Standard ‘nucleotide-nucleotide BLAST [blastn]’’ from the list underneath the
heading “Nucleotide BLAST”.
Transfer RNA molecules have a complex tertiary structure that is essential for their
function. It is dependent on intramolecular complementary base pairing and can be seen
in the diagram below of the crystal structure of a phenylalanyl tRNA from yeast. The
requirement that tRNAs maintain this structure in order to interact with the ribosome and
with aminoacyl tRNA synthetases results in strong selection in favor of retaining the
primary sequences through evolution.
Crystal structure of phenylalanyl tRNA from yeast (MMDB 14286).
Page 1 of 6
© 2002 by W. H. Freeman and Company
Exploring Genomes: Using BLAST to Compare Nucleic Acid Sequences
By Paul G. Young
Figure 3.21 (below) from your text shows the cloverleaf stem-loop secondary structure of
a tRNA molecule. This conserved structure is common to all tRNA molecules. In the
diagram, we can more clearly see the complementary base pairing in the stems.
Ribosomal RNAs that serve a structural role in the ribosome also show strong
conservation of some regions. As we will see later, however, the constraints on the
structure of messenger RNA sequences are less and there is correspondingly less
conservation of sequence.
Figure 3-21 from Modern Genetic Analysis, Second Edition.
Page 2 of 6
© 2002 by W. H. Freeman and Company
Exploring Genomes: Using BLAST to Compare Nucleic Acid Sequences
By Paul G. Young
BLASTN is very similar to BLASTP but obviously will only accept nucleic acids as
input sequence. Cut and paste the phenylalanyl tRNA from Drosophila melanogaster into
the ‘Search’ text box.
>gi|174308|gb|K00349.1|DROTRF2 D.melanogaster phe-trna-2
Just as for BLASTP, the program is run by pressing the ‘BLAST!’ button and then the
‘Format!’ button when it appears.
A search against the entire database at GenBank yields many hits, since tRNA molecules
are highly conserved. The output list from this search shows many strong hits with low E
values. Not surprisingly, many are from Drosophila—but also note other species,
including human, in the list.
Look at the alignment for one of the human subject sequences by clicking on the Score
‘137’. You should see that it is almost identical to the query sequence.
Now let’s look at the source for many of these sequences by clicking on the GI or
accession number of any of the sequences. You will find that they are mostly in very
large genome sequence files containing multiple genes; the tRNA subject sequence is
somewhere internally in the DNA fragment.
Let’s look at more distantly related hits, which would be much further down the list on
the previous page. To illustrate a hit with low identity, first you should close the window
that opened with the first search results in it. Then, click ‘Nucleotide’ on the nucleotidenucleotide BLAST page, so that you can start a new search.
Now, search again using our Drosophila tRNA, but choose the E. coli database option.
>gi|174308|gb|K00349.1|DROTRF2 D.melanogaster phe-trna-2
Then, press ‘BLAST!’ and then ‘Format!’
Page 3 of 6
© 2002 by W. H. Freeman and Company
Exploring Genomes: Using BLAST to Compare Nucleic Acid Sequences
By Paul G. Young
The output of our search is similar to the previous one—but note the E values. At this
evolutionary distance (Drosophila to E. coli), there are only a few tRNAs retaining
substantial similarity to the query sequence. The graphic display shows that most of these
hits are towards the 5’ end of our query sequence.
Click on the ‘Score’ for the top subject sequence to see the alignment. Note that it is for
the first 25 nucleotides of the query and for nucleotides 8811-8835 of the subject.
Remember that BLAST is a local alignment tool that finds high-scoring regions within a
comparison. It does not usually produce an alignment from one end of the gene to the
The subject file is for a long genomic sequence. Note in the information line that there
are 400 such segments in the database representing the entire E. coli genome. Our hit is in
genome segment 22. At position 8811-8835 within the large sequence fragment is a
tRNA sequence for an E. coli threonyl-tRNA.
Click on the GI or accession number to access the Entrez data for this large genome
sequence in E. coli.
Scroll down the data until you see the link for the gene at position 8811-8886. (You will
find it near the bottom of the data, just above the code for the entire sequence.) Now
click on the word ‘gene’ on the left of the position 8811-8886.
This link takes us to the Sequence feature view of the region. Scroll down to the bottom
of the data, and you will see the actual sequence for this particular region.
BLAST does not attempt to do a global alignment from one end of the sequence to the
other unless there is an uninterrupted high score throughout the alignment. Our hit
reported a match of 23/25 for the 5’end of the molecule. If we align the complete
sequence, we see that there are blocks of identity throughout the molecule with a match
of 51/76 but only the block from 1-25 gives a high enough score to be reported by the
BLAST program.
Remember the E value for this hit (0.015). This is the probability of finding a match of
23/25 in the database (E. coli genome) by chance alone. It is very high because the
complexity of nucleic acids is low, having only the four possible base choices at each
position. An alignment of two proteins with 23/25 identities would have a much lower E
value because there are 20 potential choices at each position and the chance of finding the
corresponding matches would be low.
Page 4 of 6
© 2002 by W. H. Freeman and Company
Exploring Genomes: Using BLAST to Compare Nucleic Acid Sequences
By Paul G. Young
By examining the full sequence as shown below, it is obvious that the subject has
significant similarity over its full length.
Now that we have had a look at highly conserved tRNA sequences, let’s try mRNA.
Messenger RNA sequences for the same conserved protein drift substantially through
evolutionary time. In addition to the drift of the amino acid sequence over time, the
degeneracy of the genetic code may have caused the same amino acid to be encoded in
multiple ways. Moreover, since the tertiary structural constraints are not as large on
mRNA as they are on tRNA or ribosomal RNA, the sequences can diverge quite rapidly.
The p34 cyclin-dependent protein kinase is essential for mitosis and is strongly conserved
at the amino acid level throughout the eukaryotic world.
Close the window with the E. coli results in it, and then click ‘Nucleotide’ from the top
of the previous search page.
Now enter the accession number ‘M12912’ (for the sequence for the p34 gene from S.
pombe, the fission yeast) into the Search box, and click ‘BLAST!’ and then ‘Format!’
When the results are in, we get a few strong subject hits. The graphic and hit list show
strong alignments to several GenBank records for the fission yeast p34 gene itself. Not
surprising, they are identical to each other. After that, however, the hits are only for very
small regions, scattered throughout the gene and with significant scores in only a few
other organisms, mostly other fungi. They do however show clustering in some
particular regions presumably associated with conserved domains essential for function
of protein kinases.
For comparison, if we run a BLASTP search using the protein product of the fission yeast
p34 gene (accession P04551), we would get a very different result.
Close the window with the S. pombe nucleotide results in it, and then click ‘Protein’
from the top of the previous search page.
Now enter the accession number ‘P04551’ into the Search box, and click ‘BLAST!’ and
then ‘Format!’
Page 5 of 6
© 2002 by W. H. Freeman and Company
Exploring Genomes: Using BLAST to Compare Nucleic Acid Sequences
By Paul G. Young
Here the hit list is extensive with very low E values and clearly we have found a large
number of homologues of this gene.
If we look at the third match (Click the score ‘408’), we see the homolog for Homo
sapiens. Here you can see that the protein in the human genome is 66% identical and
80% positive (identities and similarities) to that of the yeast. Furthermore, the similarity
extends over the full length of the protein.
There is a lesson in this. If you are searching for matches to a protein encoding sequence,
then always search with the translated product.
Now that you have experience doing nucleotide and protein searches, do a search
yourself using a gene that is mentioned somewhere in your textbook. Find the accession
numbers for the DNA sequence and for the protein sequence in Entrez (make sure both
are for the same species and that neither is for a multiple sequence file). Run BLASTN
and then run BLASTP and analyze the result.
Page 6 of 6
© 2002 by W. H. Freeman and Company