Take the 2-minute tour ×
Biology Stack Exchange is a question and answer site for biology researchers, academics, and students. It's 100% free, no registration required.

I've worked pretty frequently with genome and transcriptome data for several years now, but I'm still not 100% sure I understand the proper usage for certain terminology related to transcripts and transcriptomics. Let me give it my best shot.

  • cDNA: in general, this term corresponds to any DNA that is derived from an RNA template via reverse transcription (complementary DNA); however, it is often used to refer to full length cDNAs, DNA sequences that correspond to complete mRNA molecules
  • EST: expressed sequence tags correspond to cDNA (sub)sequences obtained from a single read of a cDNA sequencing experiment
  • RNA-seq: this refers to the use of high-throughput sequencing technologies to characterize entire transcriptomes, such as by transcript assembly or expression analysis

Here's my concern: I've seen ESTs described as "one-shot" or "single-pass" sequences, which I interpret to mean a single sequence read. However, they are also typically described as being hundreds of base pairs in length (Wikipedia claims about 500-800bp in length max), which (currently) limits this definition to reads derived from Sanger-style sequencing.

Does this mean that a short read derived from a high-throughput sequencing platform is not an EST? Is this distinction based on read length? platform? throughput?

Due to technical limitations, short reads are often assembled before they are used for other downstream analysis. I have sometimes heard the contigs belonging to these assemblies as ESTs, but this doesn't seem to be consistent with the previous definitions.

Is our terminology becoming inconsistent due to the rapid improvements in sequencing technology, or am I missing or misinterpreting something?

share|improve this question
    
Although for historical reasons an EST is usually a longish sequence (ESTs were defined in 1991 when only Sanger sequencing was available so reads would naturally be several hundred bases in length), operationally it really only needs to be long enough to tag an expressed sequence, i.e. to provide evidence for the existence of an mRNA from a specific region of a genome. –  Alan Boyd Sep 19 '12 at 19:09

2 Answers 2

My 2 cents:

To answer your main question:

Is our terminology becoming inconsistent due to the rapid improvements in sequencing technology, or am I missing or misinterpreting something?

I think what you are witnessing (being subject to(?)) is, as you say, an inconsistency of terms due to the rapid improvements in the field.

I put the caveat in that you are being subjected to this because I suspect the people referring to the "things" that a second-generation-sequencing protocol generates as "ESTs" are using a familiar word to describe something that might be unfamiliar to them.

Does this mean that a short read derived from a high-throughput sequencing platform is not an EST? Is this distinction based on read length? platform? throughput?

I guess it's (i) construction; (ii) platform; (iii) throughput; and (iv) relative "quantitativeness".

Due to technical limitations, short reads are often assembled before they are used for other downstream analysis.

You lost me here with the word "often" -- are you talking about RNAseq now or is this an EST. If you have a reference genome, I think assembling RNAseq reads before downstream analysis (alignment(?)) is an exception instead of a rule ... which is why I'm lost.

share|improve this answer
    
Ok, perhaps my use of "often" is biased by my own experience. I am working on genome annotation, and both the Maker annotation pipeline and some of our in-house workflows require RNA-seq data to be assembled before they can be used as evidence when predicting gene models. I understand of course when mapping short reads to a genome or quantifying gene expression, assembly is not a prerequisite. –  Daniel Standage Sep 19 '12 at 17:08
    
I think RNA-Seq assembly prior to alignment is common in non-model species, where you need to first describe the features you align the reads to. –  Richard Smith-Unna Sep 20 '12 at 15:56

The quick answer is that you are misinterpreting ESTs. EST stands for Expressed Sequence Tag. They are obtained by collecting and sequencing any mature (i.e. poly-A) mRNAs found in a living cell. Well, it was living until the experiment was performed at any rate :).

The term EST is not applicable to sequences derived from genomic data. It only applies to expressed sequences. You do have ESTs that are derived from cDNA libraries but these are expressed cDNAs. Short reads are not ESTs because we have no information on whether that sequence is actually expressed in a living cell.

You could apply the term to RNAseq reads (depending on the experimental procedure involved) but that is not what most people would understand by the term.

In summary, the operative term of EST is "expressed". You cannot call anything an EST if you do not have evidence that it is actively expressed in a living cell.

share|improve this answer
    
When I said "short reads", I was referring in that case to RNAseq reads. Of course, in general short reads can be genomic (if you're sequencing the genome) or transcriptomic (if you're doing RNAseq analysis or something similar). –  Daniel Standage Sep 20 '12 at 17:33
    
Still, RNAseq can also pick up non-poly A RNAs. Personally, I would only use "ESTs" to describe sequences derived from "naturally occurring" mRNAs. Given the recent findings of the ENCODE project, most of the genome seems to be actively transcribed. Such transfrags do not, again in my personal opinion, qualify as ESTS. –  terdon Sep 20 '12 at 18:28
    
@terdon: I must be missing some subtlety when you qualify ESTs as having been "found in a living cell" ... RNA-seq libraries are constructed from "living cells" as well, so what's the difference here? Capturing these reads are evidence enough that they are expressed (noise control required, obviously) Also: I suspect "mature mRNAs" are captured for "EST-ifying" much the same way they are captured for constructing a highthroughput sequencing library (via oligo-dTs to prime the polyA tail ... or even "cap trapping" if you're doing some *CAGE experiment). Can you clarify those two points a bit? –  Steve Lianoglou Sep 21 '12 at 13:31
    
@SteveLianoglou, I was differentiating more between cDNA libraries made from DNA templates and ESTs. Yes RNAseq reads come from living cells. In my answer I just say that "most people" would not call the ESTs. Not that they are not indicative of expression. I quite agree that technically you could call RNAseq reads ESTs, its just that the term is more commonly used for the older EST type data. –  terdon Sep 21 '12 at 14:09

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.