BLAST: Basic Local Alignment Search Tool Explained

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-11

Key Takeaways

• BLAST is a heuristic algorithm that prioritizes speed over guaranteed optimality by employing a seeding-and-extension strategy to identify local sequence similarities, making it efficient for large-scale database searches.
• The selection of the appropriate BLAST program (e.g., blastn for nucleotide-nucleotide, blastp for protein-protein, blastx for nucleotide-translated-to-protein, tblastn for protein-to-nucleotide-translated, tblastx for translated-translated) is critical and depends on the query and database sequence types and the biological question.
• Statistical significance is quantified by bit scores (database-independent measure of alignment quality) and E-values (database-size-dependent expectation of chance alignments), which must be interpreted in conjunction to assess the biological relevance of a match.
• Effective sequence analysis necessitates proactive handling of biological complexities such as low-complexity regions and repetitive elements through filtering and masking to prevent spurious high-scoring alignments that obscure true homologs.
• While BLAST is foundational for tasks like ortholog detection (often via reciprocal best hits) and domain identification, more sensitive methods like PSI-BLAST or HMMER are required for identifying distantly related sequences or weak evolutionary signals.

Introduction to BLAST

What is BLAST?

The Basic Local Alignment Search Tool (BLAST) is a family of algorithms and software programs designed to compare primary biological sequence information—nucleotide or amino acid sequences—against a database of known sequences. BLAST identifies regions of local similarity between a query sequence and subject sequences in a database, statistically evaluates the significance of those matches, and returns a ranked list of alignments. The fundamental principle underlying BLAST is that functional or evolutionary relationships between sequences manifest as conserved regions of primary structure, and these regions can be detected without requiring global alignment across the entire sequence length.

BLAST is not a single program but a suite of tools tailored to different sequence types and comparison strategies. The most commonly used variants—blastn, blastp, blastx, tblastn, and tblastx—differ in whether they compare nucleotide or protein sequences and whether translation is performed on the query, the database, or both. The choice of program depends on the nature of the query sequence and the type of database being searched.

The algorithm's design prioritizes speed without sacrificing sensitivity. Unlike exhaustive dynamic programming approaches such as the Smith-Waterman algorithm, which guarantees optimal local alignments but is computationally prohibitive for large database searches, BLAST uses a heuristic approach that first identifies short, high-scoring seed matches and then extends them. This strategy reduces the search space dramatically while retaining the ability to detect biologically meaningful similarities.

Historical context and development

BLAST was developed by Stephen Altschul, Warren Gish, Webb Miller, Eugene Myers, and David Lipman at the National Center for Biotechnology Information (NCBI), with the foundational paper published in the Journal of Molecular Biology in 1990. The algorithm was a direct response to the computational bottleneck posed by earlier sequence comparison tools such as FASTA, which, while faster than full dynamic programming, still required substantial computational resources for large-scale database searches.

The original BLAST algorithm introduced the concept of using short "words" (typically 3 amino acids for protein searches and 11 nucleotides for DNA searches) as seeds. These seeds are scored using substitution matrices—initially PAM matrices, later replaced by BLOSUM matrices—to identify statistically significant matches. The 1990 paper also formalized the statistical framework for evaluating alignment significance, introducing the E-value (expectation value) as a measure of the number of alignments with a given score that would be expected to occur by chance in a database search.

Subsequent refinements have expanded BLAST's capabilities. Gapped BLAST, introduced in 1997, allowed for the introduction of gaps in alignments, substantially improving sensitivity for distantly related sequences. PSI-BLAST (Position-Specific Iterated BLAST) added an iterative profile-based search strategy that detects weak but biologically meaningful similarities. The current NCBI BLAST+ suite, released in 2009, provides a modular command-line interface with improved performance and maintainability. The algorithm remains the de facto standard for sequence similarity searching, with millions of searches performed daily against the NCBI's non-redundant (nr) database.

The BLAST Algorithm

Seeding and word matching

The BLAST algorithm operates in three principal stages: seeding, extension, and statistical evaluation. The seeding stage begins by compiling a list of all short subsequences (words) of a fixed length w from the query sequence. For protein searches, w is typically 3 amino acids; for nucleotide searches, w is typically 11 nucleotides. Each word is then scored against all possible words of the same length using a substitution matrix—BLOSUM62 for proteins, or a simple match/mismatch scoring scheme for nucleotides.

Words with a score exceeding a threshold T are retained as "neighborhood words." The threshold T is a critical parameter: a low threshold increases sensitivity (more seeds are retained) at the cost of speed, while a high threshold reduces the number of seeds and accelerates the search but may miss weak similarities. For protein searches with BLOSUM62, a typical threshold is T = 11, meaning that only words with a summed score of at least 11 against the query word are kept. This neighborhood word generation step is what distinguishes BLAST from simpler k-mer matching approaches: it allows for conservative substitutions in the seed region, not just exact matches.

The database is then scanned for occurrences of these neighborhood words. This scanning step is highly optimized using a lookup table or a finite-state automaton, allowing the algorithm to rapidly identify positions in database sequences where a seed match occurs. Each such position represents a potential alignment start point, or "hit." The number of hits is typically far smaller than the total database length, which is the source of BLAST's computational efficiency.

Extension and gap handling

Once a hit is identified, the algorithm attempts to extend the alignment in both directions from the seed, without gaps in the original implementation. Extension continues as long as the alignment score increases or does not decrease below a threshold. When the score begins to decline, extension stops. The original BLAST used a "two-hit" method: two seed matches on the same diagonal within a short distance were required before extension was initiated, reducing the number of spurious extensions. Gapped BLAST, introduced in 1997, modified this strategy by using a single seed match and performing a more computationally intensive gapped extension.

Gapped extension uses a dynamic programming approach restricted to a narrow band around the seed. The algorithm allows for the introduction of gaps (insertions or deletions) in the alignment, with gap opening and gap extension penalties applied. Typical parameters for protein alignments are a gap opening penalty of 11 and a gap extension penalty of 1 (in BLOSUM62 score units). For nucleotide alignments, the gap opening penalty is typically 5 and the gap extension penalty is 2. These penalties are applied as negative scores, and the extension proceeds until the accumulated score drops below a cutoff.

The extension step is computationally expensive relative to seeding, but because it is only performed for a limited number of seed hits, the overall cost remains manageable. The result of extension is a high-scoring segment pair (HSP)—a local alignment without gaps (in the original algorithm) or with gaps (in gapped BLAST) that represents a candidate region of similarity between query and subject.

E-values and bit scores

The statistical significance of each HSP is evaluated using extreme value distribution (EVD) statistics, which model the distribution of alignment scores expected from random sequences of the same composition. The key output parameters are the bit score and the E-value.

The bit score (S') is a normalized version of the raw alignment score, calculated as:

S' = (λS − ln K) / ln 2

where S is the raw score, and λ and K are statistical parameters that depend on the scoring system and sequence composition. The bit score is independent of the database size and scoring matrix, making it comparable across different searches. A bit score of approximately 50 bits for a protein alignment typically indicates significant similarity, though this threshold varies with the search context.

The E-value is the expected number of alignments with a score equal to or better than the observed score that would occur by chance in a database of a given size. It is calculated as:

E = Kmn e^(−λS)

where m is the query length and n is the database length. The E-value is directly proportional to database size: searching a larger database yields larger E-values for the same alignment score. An E-value of 0.05 is often used as a threshold for significance, meaning that one would expect 0.05 such alignments by chance in a database of that size. For genome-scale analyses, more stringent thresholds (E < 10⁻⁵ or lower) are typically employed.

The relationship between bit score and E-value is monotonic: higher bit scores correspond to lower E-values. However, the E-value incorporates database size, so the same alignment will have a different E-value depending on the database searched. This is a common source of confusion and is addressed further in the interpretation section.

BLAST Variants and Programs

Nucleotide vs. protein BLAST

The two foundational BLAST programs are blastn and blastp. blastn compares a nucleotide query against a nucleotide database. It uses a simple scoring scheme: typically +2 for a match, −3 for a mismatch, with gap penalties of 5 (opening) and 2 (extension). blastn is optimized for finding exact or near-exact matches, such as identifying a gene within a genome, checking for vector contamination, or comparing closely related species. Because nucleotide sequences evolve more rapidly than protein sequences at the amino acid level (due to synonymous codon changes), blastn is less sensitive for detecting distant evolutionary relationships.

blastp compares a protein query against a protein database. It uses a substitution matrix—most commonly BLOSUM62—to score amino acid matches and mismatches based on the observed frequencies of substitutions in aligned protein families. The BLOSUM62 matrix was derived from blocks of aligned sequences with at least 62% identity, and it performs well for detecting moderate to distant relationships. For more divergent sequences, BLOSUM45 or BLOSUM50 may be more appropriate; for closely related sequences, BLOSUM80 or BLOSUM90 can provide better discrimination.

The choice between blastn and blastp depends on the biological question. If the goal is to identify an ortholog of a known protein in a newly sequenced genome, blastp is the appropriate choice because protein sequences retain detectable similarity over much longer evolutionary timescales. If the goal is to locate a specific DNA sequence, such as a primer binding site or a regulatory element, blastn is appropriate. For primer design and validation, the Blast Primer Search tool provides specialized functionality.

Translated BLAST tools

Three BLAST programs perform translation of nucleotide sequences to protein sequences, enabling comparisons across the genetic code. These are blastx, tblastn, and tblastx.

blastx translates the nucleotide query in all six reading frames (three forward, three reverse) and compares the resulting protein sequences against a protein database. This is useful for analyzing anonymous genomic DNA or EST sequences, where the coding potential and reading frame are unknown. blastx can detect protein-coding regions even when the nucleotide sequence contains errors or when the query is from a species with a different codon usage pattern. It is also valuable for identifying genes in newly sequenced genomes before gene prediction algorithms have been applied.

tblastn translates a protein query against a nucleotide database that is translated in all six reading frames on the fly. This is the tool of choice when searching a genome assembly for a protein of interest, particularly when the gene structure (intron-exon boundaries) is unknown. Because tblastn searches translated nucleotide sequence, it can detect protein-coding exons even when they are interrupted by introns in the genomic sequence. The six-frame translation of the database is computationally intensive, but the algorithm's seeding strategy limits the number of translations that must be performed.

tblastx translates both the query and the database in all six reading frames and compares the resulting protein sequences. This is the most sensitive translated search but also the most computationally expensive. tblastx is typically used for comparing two genomes or transcriptomes to identify conserved coding regions, particularly between distantly related species where nucleotide-level similarity has been eroded but protein-level similarity remains detectable.

The following table summarizes the BLAST programs and their applications:

Interpreting BLAST Results

Understanding E-values and bit scores

The E-value and bit score are the two most important statistical measures in BLAST output, and they must be interpreted together. The bit score is a normalized measure of alignment quality that is independent of database size. It reflects the raw alignment score after correction for the statistical parameters of the scoring system. A bit score of 50 or higher for a protein alignment is generally considered significant, but this threshold is heuristic and depends on the query length and the evolutionary distance being probed.

The E-value contextualizes the bit score by accounting for the size of the database searched. An alignment with a bit score of 50 might have an E-value of 0.001 when searching a small database of 10,000 proteins, but the same alignment would have a much larger E-value (e.g., 0.5) when searching the nr database containing over 500 million proteins. This is because the E-value scales linearly with database size: the larger the database, the more likely it is that a given score will occur by chance.

A common mistake is to interpret E-values as probabilities. The E-value is an expectation—the number of alignments with a given score expected by chance—not a probability. An E-value of 5 means that five alignments with that score or better are expected by chance in a database of that size. This does not mean the alignment has a 5% probability of being a false positive; the interpretation depends on the context. For a single query against a large database, an E-value of 5 might be acceptable for exploratory purposes, but for high-confidence annotation, E-values below 10⁻⁵ are typically required.

Reading pairwise alignments

The pairwise alignment section of a BLAST result displays the query and subject sequences aligned with a middle line indicating identities (matching residues) and conservative substitutions (represented by a plus sign in protein alignments). Gaps are shown as dashes. The alignment is flanked by coordinates indicating the positions in the query and subject sequences.

Several features of the alignment deserve attention. The "frame" field (for translated searches) indicates the reading frame of the query and subject. The "identities" field reports the percentage of identical residues in the aligned region, while the "positives" field reports the percentage of residues with positive substitution scores (including identities and conservative substitutions). The "gaps" field reports the percentage of gap positions in the alignment.

The alignment length is often shorter than the full query length, reflecting the local nature of the alignment. A high-scoring alignment that covers only a small fraction of the query may indicate a shared domain rather than a full-length ortholog. Conversely, a lower-scoring alignment that spans nearly the entire query length may be more biologically meaningful for inferring orthology. The coverage (the fraction of the query that is aligned) should always be considered alongside the E-value and bit score.

BLAST Databases and Search Strategies

Choosing the right database

The choice of database is as important as the choice of BLAST program. The NCBI maintains several standard databases, each with distinct characteristics. The non-redundant (nr) database is the most comprehensive protein database, containing sequences from GenBank, RefSeq, PDB, SwissProt, PIR, and PRF, with redundant sequences removed. As of 2024, nr contains over 500 million protein sequences. Searching nr is appropriate when the goal is to identify any known homolog of a query sequence, but the large size increases E-values and computational time.

RefSeq (Reference Sequence) databases are curated, non-redundant collections of genomic DNA, transcripts, and proteins. RefSeq proteins are annotated with functional information and are ideal for identifying well-characterized homologs. The RefSeq genomic database is organized by organism, allowing targeted searches against specific species or taxonomic groups.

For metagenomic or environmental samples, the env_nr and env_nt databases contain sequences from environmental sequencing projects. The patnr database contains patented sequences. For taxonomic-specific searches, NCBI allows restricting searches to a particular organism or taxonomic group using the "organism" field in the search interface.

The choice of database affects both sensitivity and specificity. Searching a smaller, curated database (e.g., SwissProt, which contains only manually annotated proteins) reduces the number of spurious hits and simplifies interpretation, but may miss homologs that are only present in larger, less curated databases. A common strategy is to first search a curated database for high-confidence hits, then search nr for more distant homologs.

Optimizing parameters for sensitivity vs. speed

BLAST parameters can be adjusted to balance sensitivity against computational speed. The most important parameters are the word size (w), the threshold T, and the scoring matrix.

For protein searches, reducing the word size from the default of 3 to 2 increases sensitivity for detecting short, conserved motifs but increases search time substantially. The threshold T can also be lowered to retain more seeds. For nucleotide searches, reducing the word size from 11 to 7 increases sensitivity for short or divergent sequences, which is useful for primer design and for detecting short conserved regulatory elements. The Blast Ncbi Search interface exposes these parameters in the "Algorithm parameters" section.

The scoring matrix for protein searches should be chosen based on the expected evolutionary distance. BLOSUM62 is the default and works well for most searches. For closely related sequences (e.g., within a genus), BLOSUM80 provides better discrimination. For distantly related sequences (e.g., across phyla), BLOSUM45 is more sensitive. The gap penalties should be adjusted to match the matrix: BLOSUM62 typically uses gap opening 11 and gap extension 1; BLOSUM45 uses gap opening 15 and gap extension 2.

The "expect threshold" (E-value cutoff) determines which alignments are reported. Lowering the threshold (e.g., to 0.001) reduces the number of reported hits and increases specificity. Raising the threshold (e.g., to 10) increases sensitivity but also increases the number of false positives. For exploratory searches, a threshold of 10 is reasonable; for annotation, a threshold of 10⁻⁵ or lower is recommended.

BLAST in Phylogenetics and Molecular Evolution

Ortholog detection

BLAST is a foundational tool for identifying orthologous genes—genes in different species that descend from a single gene in the last common ancestor. The standard approach is reciprocal best hit (RBH) analysis: gene A in species 1 is searched against the proteome of species 2, and the best hit (gene B) is identified. Gene B is then searched back against the proteome of species 1. If gene A is the best hit of gene B, the two genes are considered putative orthologs.

RBH analysis using blastp is computationally efficient and works well for closely related species. However, it has limitations. Gene duplication events can complicate orthology inference: if a gene has undergone duplication in one lineage, the reciprocal best hit may be a paralog rather than the true ortholog. More sophisticated approaches, such as OrthoMCL or InParanoid, use BLAST to identify all pairwise similarities and then cluster them into ortholog groups using graph-based algorithms. These methods account for lineage-specific duplications and losses.

The Gene Duplication process is a major source of complexity in ortholog detection. When a gene is duplicated, the two copies may undergo subfunctionalization or neofunctionalization, leading to divergent evolutionary rates. BLAST-based approaches can detect the similarity between paralogs, but distinguishing orthologs from paralogs requires phylogenetic analysis. For this reason, BLAST is typically used as a first-pass filter, with candidate orthologs subsequently verified by constructing gene trees and reconciling them with species trees.

Domain and motif identification

BLAST can identify conserved domains and motifs within proteins, providing insights into protein function and evolutionary relationships. When a query protein is searched against a database, the alignments often reveal that different regions of the query match different database sequences. This pattern indicates the presence of multiple domains, each with its own evolutionary history.

For systematic domain identification, the Conserved Domain Database (CDD) at NCBI uses RPS-BLAST (Reverse Position-Specific BLAST) to search query sequences against position-specific scoring matrices (PSSMs) derived from domain families. This approach is more sensitive than standard blastp for detecting domains because it uses profile information rather than a single consensus sequence. The results are displayed graphically, showing the domain architecture of the query protein.

For de novo motif discovery, PSI-BLAST is a powerful tool. PSI-BLAST performs iterative searches: the first iteration uses a standard blastp search, and the resulting alignments are used to construct a position-specific scoring matrix (PSSM). The PSSM is then used to search the database again, detecting more distant homologs. The process is repeated until convergence or a specified number of iterations. PSI-BLAST can detect very weak similarities that are invisible to a single blastp search, making it valuable for identifying novel members of protein families. However, the iterative process can be contaminated by spurious hits, so the results should be carefully inspected.

The detection of conserved domains and motifs is often the first step in functional annotation. Once a domain is identified, the Gene Ontology Online Tool and Gene Ontology Analysis Tool can be used to associate the domain with functional categories such as molecular function, biological process, and cellular component.

Common Pitfalls and Best Practices

Low-complexity filtering

Low-complexity regions—sequences with biased amino acid or nucleotide composition, such as proline-rich repeats, alanine tracts, or AT-rich stretches—produce spurious high-scoring alignments in BLAST searches. These regions are not evolutionarily informative and can dominate the search results, obscuring genuine homologs. The standard remedy is to enable low-complexity filtering, which masks these regions with the letter "X" (for proteins) or "N" (for nucleotides) before the search.

The default BLAST settings enable low-complexity filtering using the SEG algorithm for proteins and the DUST algorithm for nucleotides. However, filtering can mask biologically meaningful regions. For example, zinc finger domains contain cysteine and histidine residues that may be flagged as low-complexity. If the query is a zinc finger protein, filtering may remove the most conserved regions. In such cases, it may be necessary to disable filtering or to use a more targeted approach, such as searching with only the domain region.

Handling repetitive sequences

Repetitive elements, such as transposons, retrotransposons, and satellite DNA, present a significant challenge for BLAST searches. These elements are present in multiple copies throughout a genome, and a query containing a repetitive element will produce hundreds or thousands of hits with high scores, many of which are biologically uninformative. The Transposable Element content of a genome can be substantial—over 50% in some plant genomes—so this is not a rare problem.

The standard approach is to mask repetitive elements before searching. The RepeatMasker program identifies and masks known repetitive elements using a library of consensus sequences. Alternatively, the NCBI BLAST interface provides an option to filter low-complexity regions, which partially addresses the problem. For genome-scale analyses, it is advisable to mask repeats before performing BLAST searches to avoid wasting computational resources on spurious hits.

When to use alternative tools

BLAST is not always the optimal tool for sequence similarity searching. For detecting distant homologs, profile-based methods such as HMMER (which uses hidden Markov models) and PSI-BLAST are more sensitive than standard BLAST. HMMER is particularly effective for searching a query sequence against a database of protein families, such as Pfam. The choice between BLAST and HMMER depends on the question: BLAST is faster and more appropriate for finding close homologs, while HMMER is more sensitive for detecting membership in a known family.

For comparing entire genomes or proteomes, alignment-free methods based on k-mer frequencies may be more appropriate. These methods are computationally efficient but provide less detailed information about the location and structure of similarities. For phylogenetic inference, BLAST is typically used to identify orthologs, which are then aligned using multiple sequence alignment tools and analyzed using maximum likelihood or Bayesian methods. The Neutral Theory of Molecular Evolution provides the theoretical framework for interpreting the evolutionary significance of sequence differences.

Another common pitfall is using BLAST to compare sequences of different types without considering the genetic code. For example, comparing a mitochondrial gene to a nuclear database requires awareness of the mitochondrial genetic code, which differs from the standard code at several codons. The NCBI BLAST interface allows specifying the genetic code for translated searches, and this should be set correctly to avoid spurious results.

Summary and Further Resources

Online tutorials and documentation

The NCBI BLAST help documentation provides comprehensive information on algorithm parameters, database descriptions, and output formats. The BLAST tutorial on the NCBI website offers step-by-step guidance for common search scenarios. The EMBL-EBI provides similar resources for its BLAST implementation. For advanced users, the BLAST+ command-line tools are documented in the NCBI BLAST+ manual, which covers installation, configuration, and scripting. The Codon Optimization Tool Free resource is useful for reverse-translating protein sequences for expression studies, which is a common downstream application of BLAST-based gene identification.

Frequently Asked Questions

What is the basic local alignment search tool (BLAST)?

The Basic Local Alignment Search Tool (BLAST) is a family of algorithms and software programs for comparing primary biological sequences—nucleotide or amino acid—against databases of known sequences. BLAST identifies regions of local similarity, evaluates their statistical significance, and returns ranked alignments. It is the most widely used tool for sequence similarity searching in molecular biology.

How does BLAST work?

BLAST works in three stages. First, it generates a list of short "words" from the query sequence and identifies database sequences containing matching or similar words (seeding). Second, it extends these seed matches in both directions to generate high-scoring segment pairs (HSPs), allowing for gaps in gapped BLAST. Third, it evaluates the statistical significance of each HSP using extreme value distribution statistics, reporting bit scores and E-values.

What is an E-value in BLAST?

The E-value (expectation value) is the number of alignments with a score equal to or better than the observed score that would be expected to occur by chance in a database of a given size. It is not a probability. Lower E-values indicate more significant alignments. An E-value of 0.05 is often used as a significance threshold, but more stringent thresholds are recommended for genome-scale analyses.

What is the difference between blastn and blastp?

blastn compares a nucleotide query against a nucleotide database using a simple match/mismatch scoring scheme. blastp compares a protein query against a protein database using a substitution matrix (typically BLOSUM62). blastn is appropriate for finding exact or near-exact nucleotide matches, while blastp is more sensitive for detecting distant evolutionary relationships because protein sequences are more conserved than nucleotide sequences.

When should I use blastx or tblastn?

blastx translates a nucleotide query in all six reading frames and compares the resulting proteins against a protein database. It is used for analyzing anonymous DNA or EST sequences to identify protein-coding regions. tblastn translates a nucleotide database in all six reading frames and compares a protein query against the translated database. It is used to find genes in genomic sequences, particularly when intron-exon structure is unknown.

How do I choose the right BLAST database?

The choice of database depends on the biological question. The nr database is comprehensive and appropriate for identifying any known homolog. RefSeq databases are curated and appropriate for identifying well-characterized homologs. For taxonomic-specific searches, restrict the search to a particular organism or taxonomic group. For metagenomic samples, use env_nr or env_nt. Smaller, curated databases reduce spurious hits but may miss distant homologs.

Why does BLAST sometimes miss homologous sequences?

BLAST can miss homologous sequences for several reasons. The heuristic seeding strategy may fail to detect similarities that are too divergent to share a seed match. Low-complexity regions may produce spurious hits that obscure genuine homologs. Repetitive elements may dominate the results. The choice of scoring matrix and gap penalties may be inappropriate for the evolutionary distance being probed. In such cases, more sensitive tools such as PSI-BLAST or HMMER may be necessary.

Key Takeaways

• BLAST is a heuristic algorithm that uses seeding and extension to achieve rapid local alignment, trading guaranteed optimality for practical speed.
• The choice of BLAST program (blastn, blastp, blastx, tblastn, tblastx) must match the sequence types and the biological question.
• E-values are database-size-dependent expectations, not probabilities; bit scores are database-independent and comparable across searches.
• Low-complexity filtering and repeat masking are essential to avoid spurious hits in real-world sequence data.
• Reciprocal best hit analysis using BLAST is a standard approach for ortholog detection, but gene duplication requires more sophisticated methods.
• PSI-BLAST and HMMER provide greater sensitivity for detecting distant homologs and should be used when standard BLAST fails.
• Database selection and parameter optimization are as important as the algorithm itself for obtaining biologically meaningful results.

Further Reading

• Camacho C et al. ElasticBLAST: accelerating sequence search via cloud computing. BMC bioinformatics. 2023. PubMed 36967390
• Camacho C et al. ElasticBLAST: Accelerating Sequence Search via Cloud Computing. bioRxiv : the preprint server for biology. 2023. PubMed 36789435
• Altschul SF et al. Basic local alignment search tool. Journal of molecular biology. 1990. PubMed 223171280360-2)
• Mount DW. Using the Basic Local Alignment Search Tool (BLAST). CSH protocols. 2007. PubMed 21357135
• Nowicki M, Bzhalava D, BaŁa P. Massively Parallel Implementation of Sequence Alignment with Basic Local Alignment Search Tool Using Parallel Computing in Java Library. Journal of computational biology : a journal of computational molecular cell biology. 2018. PubMed 30004240
• Dowd SE et al. Windows .NET Network Distributed Basic Local Alignment Search Toolkit (W.ND-BLAST). BMC bioinformatics. 2005. PubMed 15819992

Related Clinical & Scientific Guides

• MAPK Pathway: Mechanism, Function, and Clinical Relevance
• Mammalian Cell Culture Bioreactors: A Practical Guide
• Nucleotide Formation: Biosynthesis and Assembly of DNA/RNA Building Blocks