Probably this is very trivial but I am self-learning bioinformatics and I don't know who to ask for this kind of stuff. Also, apparently it is really hard to find information on general topics like this.

I am learning how to use BLAST. I tried to use the sequence of the COVID-19 delta variant spike protein. I used BLASTP and I obtain as results a lot of proteins (all classified as surface glycoprotein [Severe acute respiratory syndrome coronavirus 2]) that vary just for a single aminoacid. The majority of these changes is I to V or viceversa.

What does it mean? Are these all variants of the protein? Are these all sequences with mistakes? And why does it entail only the interchange between I and V?

  • $\begingroup$ It would help to know the blastp database you use (e.g. just default with the "run blast" button?), what the output is, what the position is, etc. Briefly, we would expect those to all be real variants, it is comforting that they are replicated in different isolates. $\endgroup$ Dec 7, 2021 at 20:43
  • $\begingroup$ Note: by default blastp web interface or the command-line program may only return some number of the most similar sequences ("Max target sequences" under "Algorithm Parameters" on the web interface). You may want to look into adjusting this parameter if you want more, less similar seqs. $\endgroup$ Dec 7, 2021 at 20:48
  • $\begingroup$ For the moment I just tried to get the best 50 candidates. I don't really care, it's just to try it out. I just changed the expected value to be less than 0.05. The other parameters are left to default. So from what you are saying I assume those are all genuine variants of the protein, is it correct? $\endgroup$
    – RDGuida
    Dec 7, 2021 at 21:09
  • $\begingroup$ Unless you have some reason to expect systematic error, that is my expectation. Viral protein sequences vary significantly, that is normal. Unfortunately in this case! $\endgroup$ Dec 7, 2021 at 21:16
  • $\begingroup$ However what really leaves me wondering is that (at least in the top 50 candidates) we can only see interchange of I and V. Any idea why it might be the case? $\endgroup$
    – RDGuida
    Dec 7, 2021 at 21:29

1 Answer 1


The answer is that I and V are similar in their biochemical properties. By the way I see some V-I substitutions as well as the I-V in query vs subject and also some others (e.g. B-D, L-M). These are all real and natural variations in the sequence.

If you look at the table here, you will see that Valine and Isoleucine are both hydrophobic and similar in structure. If you delve deeper into the numbers, you will see further that they have similar charges, similar sizes, and are aliphatic. This means that they are likely to conserve protein structure in the folded form. Valine is less energy expensive in synthesis for the cell, so perhaps this plays into it a bit, but as this is a virus that is hijacking the cell processes I have no real idea if the energetic cost comes into it.

BLAST, as mentioned in the comments, returns a certain number (usually 50) of the most similar sequences to the query. You can adjust the number and specificity limits during your setting of the search parameters. If you do something like blast the spike reference sequence from Wuhan-Hu-1 (YP_009724390.1) against your sequence (put yours in query, and the Wuhan-Hu-1 one in the subject or vice-versa), you will see a number of substitutions, indels etc, none of which are I-V. From this you can deduce that your findings of I-V and V-I are simply a matter of sampling based on the most similar sequences.

I don't know the algorithm for BLAST (or anything else for that matter), but it may well take similarity of biochemical properties into account when searching for most similar sequences. Given that there are millions of submitted sequences for the spike of SARS-CoV-2, there are likely to be many thousands of highly similar sequences in the database.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .