Could I know what you use BLAST for?

Question

[juanjo75es]

I would like to know what are typical use cases for a BLAST-like technology (finding local alignments on large datasets)

I know some use cases are:

• Finding homolog genes in other species
• Anotating genomes
• Predicting gene function based on similarity with other genes.

But these uses seem to be very specific and potentially useful to very few people (few people gets a new genome or a new gene to look an homolog for). And despite these uses seem not to be massive, the statistics of the NCBI site seem to indicate that BLAST is widely used with millions of queries per year.

Could I know what do you use BLAST for or what do you know it is used for?

[Joe]

The most common use case is “I have an unknown sequence, what organism is it from, or what does it look most like?” In my experience, particularly if you’re a wet lab biologist.

You say those uses are very specific, but that kind of goes without saying - blast is a specific tool for a specific job, but I think you’re vastly underestimating it when you say “it only seems useful to very few people”. It’s probably the most used bioinformatics program of all time. All of those tasks are just an extension of “what does sequence A look like?”.

Answer 1

[Jeremy Leipzig]

BLAST is a suite of alignment programs. I agree the relevance of BLASTN is not what it was during the Sanger era, but if you're looking to align a protein-coding sequence it is always better to do so in a translationally-aware (i.e. amino acids) manner - that's what the BLAST suite allows.

Program          Subject             Query
BLASTN           Nucleotide          Nucleotide
BLASTP           Protein             Protein
BLASTX           Protein             Protein
TBLASTN          Protein             Protein
TBLASTX          Protein             Protein

[Mensur Dlakic]

I don't mean to be pedantic, but BLAST is not a suite of alignment programs. Lots of people, myself included, use BLAST for alignments, but that is not its purpose. In fact, BLAST is not a good aligner once sequence identity drops below 40% or so.

BLAST is a suite of sequence comparison and scoring programs. It says so on its web page.

BLAST finds regions of similarity between biological sequences. The program compares nucleotide or protein sequences to sequence databases and calculates the statistical significance.

[Joe]

I don't really think that's a matter of opinion:

Basic Local Alignment Search Tool....

It very definitely is an alignment program.

[Mensur Dlakic]

Using the same acronym but highlighting different portion gives us:

Basic Local Alignment Search Tool ....

So yes, it is not a matter of opinion. BLAST uses local alignments to search for similar sequences. Its purpose is to search, and local alignment is the means to do it.

[Jeremy Leipzig]

The lady doth protest too much, methinks

[Mensur Dlakic]

It is endearing how you guys stand up for each other.

Answer 2

[Mensur Dlakic]

I could think of many distinct cases (>30) where I have used BLAST. It was not always the best tool for a given task, but it is almost always the fastest.

Rather than going through my exotic examples, I encourage you to think about other tools that rely on BLAST. For example, it is heavily used in proteomics. I feel confident in saying that more than 50% of machine learning applications aiming to predict something about proteins rely on BLAST. That's because BLAST has been the fastest tool for so long to make PSSM matrices (or alignment checkpoints, depending on one's preference). Even though there are tools today with comparable speed to BLAST for that purpose, they are either relatively recent or simply are not as widespread.

Answer 3

[Jean-Karim Heriche]

I use it to quickly map any kind of sequence to a reference or to quickly find/group together similar sequences. For example, I use blast to:

• map siRNAs to target transcripts
• identify/validate CRISPR/Cas9 target sites
• map mass spectrometry identified peptides to proteins
• collect similar sequences before building a multiple sequence alignment

Answer 4

[Kevin Blighe]

One of my own practical examples that I was using in South America with some colleagues: we performed RNA-seq of a bacterial strain and assembled the transcriptome to identify candidate mRNA sequences with coding potential. We then ran these sequences through blastx to assist in the interpretation of their function.

In the past, in a clinical setting in the UK NHS (National Health Service), I also downloaded a local copy of the nt BLAST database and aligned samples against it to check for contamination.

Kevin