Hello,
The RNU4ATAC gene is a non-coding gene responsible for three Mendelian diseases, therefore variant interpretation by diagnostic laboratories relies partly on the population data of gnomAD.
For non-coding genes, variant numbering depends on where the gene starts. Unfortunately, the Ensembl entry for Homo sapiens RNU4ATAC gene (ENSG00000264229) did not give the right coordinates. These wrong coordinates are used in genomAD and as a result of this mistake, the consequences of the variants are wrongly deduced, as there is a -1 shifting in the numbering of the nucleotides for all the variants.
This mistake has just been corrected in Ensembl Release 114 (May 2025): ENST00000580972.1 has been changed to ENST00000580972.2 which has the right coordinates (chr2:121,530,880-121,531,009). Could this change be implemented in gnomAD?
Thank you.
Thanks Sylvie! We are aware of this issue with RNU4ATAC and are happy that this has been addressed in the MANE transcripts for moving forward. When there are new releases of gnomAD, this should eventually get fixed but there aren’t resources at the moment to update single genes (and we wouldn’t want to introduce inconsistencies between genes).
Hello Anne,
Thank you for your answer.
I am glad you are aware of this issue with RNU4ATAC. I wrote to the address “exomeconsortium@gmail.com” 6 years ago to report this problem and never received a response, but I may have alerted someone anyway.
It took me a lot of effort to make Ensembl fix this error, so I am not going to give up now with gnomAD: a resolution of the problem at such a vague date is not satisfactory ! I understand when you say that there are no resources at the moment to update single genes, but not when you talk about the introduction of inconsistencies between genes: RNU4ATAC is the only non-coding snRNA gene whose coordinates are not accurate. Currently, there is an inconsistency that needs to be fixed ! As I was saying in my previous message, it has serious consequences for the diagnosis of patients, as for RNU4ATAC, the whole HGVS consequence column is false. Some diagnostic labs correct the mistakes (without entirely leveraging possible confusion), but others do not and interpret variants inaccurately. We are now aware of 60 different RNU4ATAC pathogenic variants identified in some hundreds of patients, so it is something that matters.
Coming back to the lack of resources to update single genes, would it take so much time and trouble to change the Ensembl canonical transcript reference from ENST00000580972.1 to ENST00000580972.2? Would that not automatically modify the HGVS consequences?
Regards,
Sylvie