1
$\begingroup$

Is there a standard way to deduce protein sequence length from gene DNA sequence length of the gene coding for it?

Naively I had assumed that amino_acid_seq_length / 3 -1 (deleting one for the stop codon)ought to work but apparently not always. Is there a better way?

Let's assume the gene is eukaryotic, specifically a plant gene.

e.g.

enter image description here

Or

enter image description here

$\endgroup$
  • 1
    $\begingroup$ (AA Seq. / 3 ) - 1 should be a good approximate in bacteria and archea. Since generally transcript RNA is mature. In eukaryotic orgsnism you have all sorts of processing complexity, so unless you intron spliced sequence, you wont know for sure the protein sequence $\endgroup$ – SciEnt Jun 26 '16 at 7:48
  • $\begingroup$ Thanks @SciEnt. But if you express a eukaryotic gene in something like E Coli the resultant recombinant protein should still be (AA / 3) - 1 since they will lack the post-processing machinery? $\endgroup$ – curious_cat Jun 26 '16 at 12:41
  • 1
    $\begingroup$ curious_cat, if so that is what we would expect, you are correct. $\endgroup$ – SciEnt Jun 26 '16 at 15:14
1
$\begingroup$

If you look at the DNA sequence in the patent, you will see that it does not start with ATG and does not end with a stop codon. The disclosed sequence has some additional bases in it, therefore the discrepancy in protein and DNA length. Those additional bases almost always occur in cDNA, e.g. because of polyadenylation, Kozak sequences, etc.

$\endgroup$
  • $\begingroup$ Thanks! So, if you were to want to use the sequence for (say) heterologous expression would it be imperative to correct those aberrations? i.e. How does one infer what the right length of the Amino Acid will be or which bases exactly are the additional bases? In other words, how does one "clean up" a cDNA sequence? $\endgroup$ – curious_cat Jun 26 '16 at 16:47
  • $\begingroup$ You can use a tool like this one: web.expasy.org/translate Just enter the sequence and look for the longest open reading frame. $\endgroup$ – Ashafix Jun 26 '16 at 16:56
  • $\begingroup$ Thanks again Ashafix. That explains a lot to me. One small question: Your tool predicts the correct 569 AA protein in the case of my second snippet. That's great. But in the case of the first sequence the tool gives 569 whereas the patent snippet lists 570. web.expasy.org/cgi-bin/translate/dna_sequences?/work/expasy/tmp/… Any idea why I'm off by one AA? Am I making a mistake? Or....? $\endgroup$ – curious_cat Jun 26 '16 at 17:07
  • 1
    $\begingroup$ Seems like patent data obfuscation to me, the actual sequence in the patent is 569 but if you look at the last 5 amino acids in the translated sequence, it is PLGEE, while in the patent amino acid sequence a aspartate appears out of nowhere: PLDEE. Two recommendations: 1) Start organizing your sequences with some program, it is easier to visually compare them 2) treat patents with caution, they have to tell the truth otherwise they will be invalid, but they don't necessarily tell the whole truth and it might be hidden deep in the patent to confuse the reader. $\endgroup$ – Ashafix Jun 26 '16 at 18:10
  • $\begingroup$ Thanks for the tips! Any recommendations for a program to organize my sequences? What do you use? $\endgroup$ – curious_cat Jun 26 '16 at 18:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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