FASTA sequences are different from PFM

[Echo Evans]

Hello JASPAR team,

I'm writing to you because I'm using your FASTA files for my research. I initially assumed that the FASTA sequences on a matrix's page were the exact ones used to generate the PFM, but I've noticed some discrepancies.

For example, with MA0020.2 (https://jaspar.elixir.no/matrix/MA0020.2/), the PFM on the main page shows that the first position is exclusively A. However, the FASTA sequences provided for this matrix clearly don't all start with A; they seem to be different.

On the other hand, for MA0303.3 (https://jaspar.elixir.no/matrix/MA0303.3/), the sequences in the provided FASTA file appear to be perfectly consistent with the PFM.

Could you please clarify what the FASTA files on the binding site information pages represent? Why is there a difference between the two cases?

Thank you very much.

[Anthony Mathelier]

Thanks for raising this issue. The FASTA file for MA0020 is indeed incorrect and should be removed as we do not have access to the original data used to construct the PFM.