What can one DNA sequence tell us about another?

Every organism carries genetic information encoded in DNA. Much of that information is shared, to varying degrees, with other organisms. When scientists discover a previously uncharacterized gene or protein, comparing its sequence with known sequences can provide the first clues about its possible biological role.

Consider a researcher who sequences a fragment of DNA from a soil bacterium. The laboratory obtains a series of nucleotide letters, but the sequence alone does not reveal which gene it represents. To investigate, the researcher searches a database containing sequences previously reported from different organisms. Some database records may share long regions of similarity with the new fragment.

Such matches can support hypotheses about shared ancestry, gene family membership or potential molecular function. However, sequence similarity is evidence that needs interpretation. Two related proteins may perform different functions, and short matches can sometimes arise by chance.

The Basic Local Alignment Search Tool, commonly known as BLAST, was developed to make these comparisons computationally practical. Its original description by Altschul and colleagues in 1990 established an influential approach to searching large sequence databases for locally similar regions.

Similarity begins with placing sequences together.

A sequence alignment places nucleotide or amino-acid positions into correspondence so their similarities and differences can be examined. Matching positions, substitutions and gaps provide information about possible relationships between the sequences.

Local alignment searches for highly similar regions within two sequences. This differs from global alignment, which attempts to align sequences across their full lengths. Local alignment is valuable when two proteins share a conserved domain but differ substantially elsewhere, or when a short DNA fragment matches part of a longer genomic sequence.

Query    A C G T A C G T
         | | | | |   | |
Subject  A C G T A T G T

Alignment scores depend on more than the number of identical letters. Nucleotide searches can assign rewards and penalties for matches and mismatches. Protein searches commonly use substitution matrices that give different scores to different amino-acid replacements. Gap-opening and extension penalties account for insertions and deletions.

Alignment is therefore a modeling decision, not simply a visual comparison. The selected scoring system and alignment method affect which similarities are reported.

How BLAST searches databases without comparing everything exhaustively

A biological sequence database can contain enormous numbers of nucleotide or protein letters. Exhaustively calculating optimal alignments between a query and every database sequence would often be unnecessarily expensive. BLAST uses heuristic search strategies to identify promising regions efficiently.

In simplified terms, a typical BLAST search identifies short sequence words associated with the query and looks for suitable matches in database sequences. Potential matches can be extended to evaluate longer regions of similarity. The program uses scoring rules and statistical models to determine which alignments deserve attention.

The exact details differ across programs and search settings. Protein searches may use neighboring words whose substitution-matrix scores exceed a threshold, whereas nucleotide searches often use exact or specially structured seed matches. Low-complexity masking can reduce misleading matches caused by repetitive sequence composition.

Heuristic searching provides useful speed, but it does not guarantee recovery of every biologically meaningful relationship under every parameter setting. Distantly related sequences can be missed, especially when similarity is weak or concentrated in short regions.

BLASTN, BLASTP, BLASTX and the different kinds of sequence searches

The BLAST family includes multiple programs designed for different combinations of nucleotide and protein sequences. Selecting the correct program is essential because the search method determines what kind of biological relationship can be detected.

BLASTN compares nucleotide queries against nucleotide databases. BLASTP compares protein sequences against protein databases. BLASTX translates a nucleotide query in possible reading frames before searching a protein database, which can help investigate potential coding regions in unfamiliar nucleotide sequences.

Imagine obtaining a DNA fragment from an organism whose genome has not been fully characterized. BLASTN may detect highly similar nucleotide sequences, while BLASTX could help reveal protein-level relationships that are less obvious from the nucleotide sequence alone.

Program choice also affects computational requirements, sensitivity and interpretation. The strongest workflow begins by identifying whether the biological question concerns DNA-level similarity, protein similarity or possible encoded proteins.

Why the choice of reference database changes the answer

A BLAST result depends on the database being searched. A match can only be found if suitable reference sequences are present and detectable under the selected search settings. Searching a narrow, curated collection answers a different question from searching a much larger and more heterogeneous database.

For example, a researcher interested in a known human gene may choose a focused reference sequence collection. A researcher investigating a previously uncharacterized environmental gene may need a broader search across sequences from many organisms.

Database content also influences statistical significance. BLAST E-values depend in part on the size of the sequence search space, so the same underlying alignment can be associated with different E-values in different searches.

Useful experimental documentation should include the query sequence or accession, selected BLAST program, database and version or retrieval date, plus relevant parameters. This makes later comparison and replication more reliable.

Understanding E-values, percent identity and query coverage

BLAST reports often present several metrics for each database match. The bit score represents a normalized alignment score. Percent identity describes the proportion of aligned positions with identical residues or bases. Query coverage indicates how much of the query is represented in the reported alignment or alignments, depending on the report field used.

The E-value, or expectation value, describes the expected number of matches with a score at least as high as the observed alignment that might occur by chance in the specified search space, under the model assumptions. Lower E-values generally indicate stronger statistical evidence that a match is not a random occurrence.

These numbers must be read together. A short alignment with high identity may cover only a small fraction of the query. Conversely, a longer alignment with moderate identity may provide stronger evidence for a relationship between full-length proteins.

The NCBI BLAST documentation explains that E-values should be interpreted in the context of search size and alignment length. A statistically strong result is useful evidence, but it does not independently prove a particular biological function.

Investigating an unfamiliar protein with BLASTP

Imagine that a researcher has identified a predicted protein sequence from a newly assembled microbial genome. The sequence is available in FASTA format, but its precise biological function remains unknown.

A useful first investigation is to search the protein sequence against a suitable reference protein database using BLASTP. The researcher reviews the ranked matches, emphasizing alignment coverage, identity, E-values and the quality of existing annotations.

Suppose several significant matches correspond to members of the same broad enzyme family. This pattern may support assignment of the new sequence to that family. However, family membership alone may not reveal the precise substrate, catalytic rate or biological role of the candidate protein.

The values above are illustrative, not results from a real BLAST experiment. They show why a match with the highest identity percentage is not automatically the most informative match. Full interpretation also requires statistical significance, domain architecture and biological evidence.

Follow-up investigation could include conserved-domain searches, review of relevant publications and, where appropriate, laboratory characterization of the protein. This illustrates how BLAST can guide experimentation rather than replace it.

Similarity does not automatically establish evolutionary history

Sequence similarity can provide important evidence of common evolutionary ancestry. Yet ranking BLAST matches by E-value or percent identity does not produce a reliable phylogenetic tree by itself.

Different genes and evolutionary lineages may evolve at different rates. Composition biases, repeated sequences and ancient substitutions can complicate the relationship between observed similarity and evolutionary distance.

If the scientific objective is to reconstruct evolutionary relationships, researchers generally need additional analyses, such as multiple sequence alignment, evolutionary model selection and phylogenetic inference. They also need to consider whether the compared genes are orthologs, paralogs or related through other evolutionary processes.

BLAST is excellent for quickly finding candidate related sequences. A rigorous evolutionary investigation asks a different question and typically requires more specialized methods.

Common BLAST interpretation mistakes and how to avoid them

One frequent mistake is treating the first result in a BLAST report as a confirmed answer. A search ranking reflects the selected scoring system, database and reporting settings. It does not establish the biological identity or function of the query beyond reasonable uncertainty.

Low-complexity sequence regions can produce statistically uninformative alignments, especially when repetitive composition creates chance similarities. Masking and carefully selected search parameters help reduce some of these problems.

Another challenge is distinguishing protein domains from complete protein relationships. Two proteins may contain one conserved domain while differing substantially in their remaining sequence. Such matches can be biologically meaningful without establishing identical overall function.

Researchers should also account for database incompleteness and annotation quality. A poorly characterized protein may resemble a predicted sequence carrying an uncertain functional label. Transferring that label without reviewing the evidence can propagate annotation errors.

Documentation supports reproducibility. Recording the database, sequence accessions, BLAST program, search date and relevant parameters allows another researcher to understand and repeat the search.

How sequence comparison accelerates biotechnology and discovery

Sequence comparison is now embedded throughout modern biological research. It supports identification of candidate genes, comparative genomics, protein-family investigation and the interpretation of newly sequenced organisms.

In microbial biotechnology, BLAST can help researchers identify sequences related to known enzymes or metabolic proteins. This allows candidate genes to be prioritized for cloning, expression and experimental testing.

In molecular biology, sequence comparison can support investigation of construct identity and possible sequence relationships. In agricultural research, comparing genes across species can help identify conserved families relevant to plant biology.

Advances in genome sequencing, protein prediction and biological databases make these comparisons increasingly valuable. At the same time, the growth of available sequences increases the importance of meaningful statistics, efficient algorithms and reliable functional annotation.

BLAST connects unfamiliar biological sequences with existing molecular knowledge. It does not replace experimental investigation; it helps researchers decide which hypotheses are worth exploring. That connection between computational evidence and biological validation makes sequence alignment a foundational tool in modern biotechnology.