I see in this paper the notation IL-1A C[-889]T which suggests that for this gene, the reference allele is C and the variant is T.

However, when I look up the same gene on dbSNP, it says the alleles are G>A / G>C / G>T.

I noticed for a number of other genes I looked up in the same paper, G becomes C and vice-versa, T becomes A and vice-versa. Is it because of different conventions regarding which DNA strand is used to represent DNA sequence (forward vs reverse)?

  • 3
    $\begingroup$ "Genetic code" refers to the translation from three-nucleotide codons to amino acids. A change in genetic code would be changing the amino acid that is associated with some triplet. $\endgroup$
    – Bryan Krause
    Commented Jan 2 at 22:52

1 Answer 1


The first is with respect to the transcriptional unit of the IL-1A gene. For example the "-" indicates that it is upstream of transcription start site.

The second is with respect to the GRCh38 build of the reference human genome, which has a defined orientation.

I don't have access to the paper that you link, but the NCBI annotation for the gene indicates that it is on the reverse strand of the reference genome sequence.

That would seem to indicate that the two annotation methods are concordant, if somewhat confusing. One refers to a position's distance from the start of the gene, one refers to a position's distance from the start of the reference chromosome. Those will necessarily diverge in the reference allele for genes transcribed on the reverse strand.

If you are interested in analyzing the behavior of a specific gene, the transcriptional unit is obviously the relevant level of analysis (for example, predicting translated protein sequence). However, if you are interested in defining global genomic coordinates (for example, mapping sequencing reads), the reference genome become the relevant level of analysis. So each is useful for different purposes.

So, in brief, the answer to your question is yes.

  • $\begingroup$ When you say "those will necessarily diverge" do you mean the reference allele will always obey the base pairing rules - like here C for the reference human genome and G when reading from the transcriptional unit ? $\endgroup$
    – Youcha
    Commented Jan 3 at 12:32
  • 1
    $\begingroup$ @Youcha Assuming that I understand correctly, then yes. C will only ever base pair with G (or at least that's a foundational assumption of how we represent genomes on computers) $\endgroup$ Commented Jan 3 at 18:44

You must log in to answer this question.

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