Introduction to Sequence Alignment 2.12 2003/01/24 2007/09/04 11:35:07.834 GMT-5 Susan Cates mscates@bioc.rice.edu Susan Cates mscates@bioc.rice.edu amino acid sequence bioinformatics BLAST BLAT nucleotide sequence pair-wise sequence alignment sequence alignment This module is an introduction to pairwise sequence alignments between a query sequence and a sequence database, using NCBI's BLAST, and between a query sequence and a genome, using UCSC's BLAT. Proteins that have a significant biological relationship to one another often share only isolated regions of sequence similarity. For identifying relationships of this nature, the ability to find local regions of optimal similarity is advantageous over global alignments that optimize the overall alignment of two entire sequences. BLAST, or Basic Local Alignment Search Tool (1), is an alignment tool that uses a measure of local similarity to score sequence alignments in such a way as to identify regions of good local alignment. The basic BLAST algorithm can be implemented in DNA and protein sequence database searches, motif searches, gene identification searches, and in the analysis of multiple regions of similarity in long DNA sequences. There are 5 BLAST options on the NCBI web site. BLASTP compares any amino acid query sequence against a protein sequence database. Similarly, BLASTN compares a nucleotide sequence against a nucleotide sequence database. BLASTX takes a nucleotide query sequence and translates it in all reading frames for comparison against a protein sequence database. TBLASTN compares a protein sequence against a nucleotide sequence database, translating the nucleotide database sequences in all possible reading frames. TBLASTX compares translations of a nucleotide query sequence against translations of a nucleotide sequence database. Sequences must be input in one of three formats; FASTA sequence format, NCBI Accession numbers, or GIs (GenBank Identifiers). The FASTA format is the only format where you can input a sequence, instead of a number or code identifying a gene that has already been deposited in GenBank, so this is the format that must be used for nonpublished sequences. Take a moment to view the FASTA format description. Consider the following nucleotide sequence, which is in FASTA format, but lacks the single-line description that identifies it source: TCAAGCAGATCACTGTCCTTCTGCCATGGCCCTGT GGATGCGCCTCCTGCCCCTGCTGGCGCTGCTGGCC CTCTGGGGACCTGACCCAGCCGCAGCCTTTGTGAA CCAACACCTGTGCGGCTCACACCTGGTGGAAGCTC TCTACCTAGTGTGCGGGGAACGAGGCTTCTTCTAC ACACCCAAGACCCGCCGGGAGGCAGAGGACCTGCA GGTGGGGCAGGTGGAGCTGGGCGGGGGCCCTGGTG CAGGCAGCCTGCAGCCCTTGGCCCTGGAGGGGTCC CTGCAGAAGCGTGGCATTGTGGAACAATGCTGTAC CAGCATCTGCTCCCTCTACCAGCTGGAGAACTACT GCAACTAGACGCAGCCCGCATGCAGNCCCCCACCC GCCGNCTTCTGCACCGAGAGAGATGGAATTAAACC CTTGAACCCAGCANANAAAAAAAAGAAATAAAA Access the NCBI BLAST home page , and click on the link for nucleotide BLAST (BLASTN), the standard BLAST for searching a nucleotide database with a nucleotide query sequence. Highlight and copy the FASTA sequence above, and paste it into the first box on the BLAST query page. BLAST allows the option of querying over only a subset of the query sequence, identified by numerical order in the "query subrange" boxes next to the query. If it were desirable to query with only the subsequence beginning with the 50th nucleotide of the example sequence and ending with the 250th nucleotide, then the number 50 would be entered into the "From" box and 250 would be entered into the "To" box. Here in this example the entire sequence should be used as the query, so these boxes should be left blank. The database can be selected under "Choose Search Set". There are many available database choices you can view from the pull down menu. The "nr/nt" database is the largest nucleotide database available through NCBI BLAST; select the "nr/nt" database for this exercise. It includes all GenBank, RefSeq Nucleotides, EMBL (European nucleotide database), DDBJ (Japanese nucleotide database) and PDB (Protein Data Bank) sequences, but no EST, STS, GSS, or phase 0, 1 or 2 htgs (unfinished high throughput genomic) sequences. The NCBI nr database originally got its name from the phrase "nonredundant" nucleotide database, but there is no longer any claim to nonredundancy in the sequence set. Choosing the largest database is not always best; in fact, often the goal is to find a match from a specific organism, for example, a lab that works with flies ( "Drosophila melanogaster" ) as the model organism might be only interested in their organism's sequences. In that case, there is a box entitled Organism, where a user interested can type in "Drosophila melanogaster" . Under program selection, the user can choose to optimize blastn for the type of sequence search that best relates to their problem, for highly similar sequences (megablast), more dissimilar sequences (discontiguous megablast)or somewhat similar sequences (blastn). At the bottom of the web page, there is a link that allows the user to view the default values for formatting (return the descriptions for the top 100 scoring alignments) and for the algorithm parameters. If you selected megablast, optimize for highly similar sequences, the word size used in the search will automatically default to 28 nucleotides. If you select blastn, optimize for somewhat similar sequences the word size will be 11 nucleotides, the historically common default value for blastn. For the purposes of this example, select to optimize for somewhat similar sequences (blastn) and click on the BLAST button to submit your query to the BLAST queue. BLAST should state that the query was successfully submitted, and provide a request ID that accesses the results. The page updates itself automatically, you do not have to refresh it. The results begin with a graphical overview that uses the query sequence as the rule, shown separately at the top of the graph. In this example, the rule (our query sequence) is 453 nucleotides in length. The score of each alignment is divided into one of 5 ranges by the five different colors in the color key provided, and the alignments are placed in the order of best scores first. The scores are normalized, and normalized scores are given in units of bits, to allow comparisons between alignments. More information on scores can and should be perused in the Altschul tutorial on the NCBI website. Clicking on any of the first 50 lines in the graph takes the user to the associated alignment, located in the last section of this results page. Look at the graph of best results at the top of the page. How long is the alignment length for the best match in the graph? The second section of the results page, immediately following the graph, contains the top 100 alignment scores and their sequence descriptions. Here, the "hits", or matched sequences returned as a list of results, contain links to their GenBank entries, their GenBank definitions under the column entitled "Accession", the normalized scores of the alignments, and the E-values. The E-values quantify the probability that the listed alignment might occur randomly. The scores are linked to the actual alignment further down the page. Clicking on a score will take you directly to the alignment that obtained that score. In some cases, there are icons containing capital letters in the links column, which are links to related entries in one of the NCBI databases, such as UniGene unique gene cluster information, Entrez Gene and GEO gene expression and hybridization array data . Look at the first 5 "hits" in the results list. What organism is the most common source of the sequences in the first 5 hits? What protein is most commonly identified in the description column of the alignments as being associated with or related to these sequences? The third section of the results page contains the actual alignments, listed in the same order as the graphed results and the scores. In the first alignment, what is the ratio of identical bases to total bases in the aligned subsequence? What is the listed percent identity? What is the score? What is the E- (Expect) value? Look for entry with the accession code BC005255.1 at the beginning of the alignment and click on the accession link. Note that this "hit" is a cDNA, or a DNA sequence that has been copied from an mRNA molecule by the enzyme reverse transcriptase. This is indicated in the description, but whenever the source of the sequence is unclear, click the link for the GI number and view the GenBank entry for more information. Look under the "COMMENT" section for additional details about this sequence. What group was responsible for the DNA sequencing for this entry? Return to the BLAST alignment results page. Scroll down to view the actual sequence alignments. The alignments are represented with vertical lines illustrating identical matches, blanks indicate no match. In these nucleotide alignments, there is no measurement for similarity between non-identical nucleotides. In protein alignments, however, chemical or structural similarity is usually identified between the amino acids. Return to the BLAST home page, and select the blastx "translated nucleotide sequence query". Cut and paste the same nucleotide sequence used in the previous example into the "Search" box. The default translation tool is BLASTX, which translates the query sequence into a protein sequence and searches a protein database. To search a translated nucleotide database with a protein query, the TBLASTN algorithm should be selected. To search a translated nucleotide database with a translated nucleotide sequence, the choice would be TBLASTX. TBLASTX is the most computationally intensive of these three searches, therefore common practice is to use BLASTX first to search protein sequence databases, and then if there are no hits with BLASTX, consider TBLASTX to translate nucleotide sequence database entries into protein sequences. For this example, use the translation algorithm BLASTX. Use the nr database and leave the default values for all other BLAST options, and hit BLAST. View the results page. What protein is listed in the vast majority of the returned matches? Look at the first alignment, which should have 100% identity between the aligned subsequences. Note that a particular alignment may have more than one sequence associated with it. Therefore, you must look at the actual alignments, not the list of scores, to answer the following questions. Why is the score lower for the BLASTX result than for the BLASTN result, even though the percent identity between aligned subsequences is 100%? How many different species are listed as sources where the aligned subsequence has an identical amino acid to total amino acid ratio of 86/86, and 100% identity? This protein sequence is highly conserved across several species, a trait that is often a sign of physiological importance. Scroll through the protein alignments. Notice that alignment between amino acids is illustrated differently than alignment between nucleotides. An identical match is shown by listing the one letter amino acid code in the middle row, between the two aligned sequences. A mismatch is indicated by a blank, but similarity is indicated by the "+" sign, meaning the amino acids are not identical, but they have some of the same chemical or structural properties. Gaps are indicated by hyphens in the sequence that contains the gaps. Gaps are penalized in an alignment, and cause the normalized score to be lower. BLAT, stands for Blast-like alignment tool (2), and is used for sequence comparisons against an entire genome. Review the difference between BLAT and BLAST at the BLAT FAQ site. Execute a BLAT search, using the sequence above, from the UCSC Genome Bioinformatics Site. Click on the link in this web page entitled "BLAT". Search the human genome, leaving the settings at the default values. The search results will first appear in summary form. Identify the sequence our query most closely matches by the highest score. What human chromosome (number) contains the sequence our query most closely aligns? What percent identity exists between our sequence and the aligned subsequence(s)? Click on the link entitled "details". First, the query sequence is presented, with matching bases capitalized in blue font. Bases in cyan mark the beginning and end of aligned subsequences, indicating a gap in either the reference or the query sequence. Second, the genomic sequence from the selected chromosome is shown, and blue, capitalized bases illustrate the matching region(s). Notice that the query cDNA sequence is missing a large region of DNA that is present in the chromosome. The pairwise alignment is shown below the genomic sequence. Return to the original results summary and click on the link entitled "browser". The missing region is illustrated in graphical form here, where the chromosome band is shown in grey, extending across the graph and the query sequence, labeled at the left as "YourSeq", is in black, below the STS Markers. In fact, the examples in this module use an EST, or an expressed-sequence tag, for the query sequence. ESTs are mRNA-derived representations of the genes expressed in a given tissue and/or at a given developmental stage. Knowing that our sequence is an EST, how could this explain the region of DNA that appears in the genomic sequence, but not in our query sequence? If the answer to this question is not intuitive, read the EST section of the NCBI primer. Scroll down to the section entitled "Genes and Gene Prediction Tracks" and change the display options under Ensembl to "pack" and the setting under GenScan to "pack". Next, scroll to the bottom of the page and hit the refresh button. (For a description of the different display options for annotation tracks, read the User Guide. The GenScan predicted genes for this area of the chromosome are shown in brown, while the Ensembl gene predictions are in maroon. How does our sequence line up with the predicted genes? Now, design an independent search in order to answer the following question. What is the GI number and definition for the EST that aligns most closely with our query sequence, according to BLAST? This has been an introduction to BLAST sequence alignment, designed to give an idea of some of the different ways to optimize searches and problems that require the use of sequence alignment tools. Subsequent modules in this course will build on this foundation, exploring advanced techniques and additional alignment tools. Altschul S.F., Gish W., Miller W., Myers E.W., Lipman D.J. Basic local alignment search tool J. Mol. Biol. 1990 215:403-410 Kent, WJ. BLAT--the BLAST-like alignment tool Genome Res. 2002 12(4):656-64