I know I can just take it from here, problem is that I am dealing with thousands of SNPs not just this one. And I'm afraid I will encounter a similar error in the rest of my files.
I first found this error using pyliftover and then when I checked on UCSC it gave the same, an error.
EDIT: this happens for several SNPs, not only this one. rs2236566 and rs12728058 are other examples.
UCSC liftover should not be used to liftover variants as it is designed for lifting over intervals. As @cmdcolin explained, there is a gap in the chain file, in part due to hg19 and hg38 representing different alleles for the very polymorphism you are trying to lift over. This is a common issue with many liftover tools as I explain in my own liftover paper. Using the BCFtools/liftover plugin this problem is automatically addressed:
In this case the region around the chain gap is automatically realigned and the correct location for the marker is then identified. The SWAP=1 INFO field explains that reference and alternate alleles were swapped
as far as why that particular position is failing to liftover, the liftover chain file has a completely replaced sequence there that might make it hard to properly covnert
see screenshot of jbrowse 2 visualization of the chain file at this position
and indeed, the NCBI shows the reference letter for the SNP changes
random note: one could argue that the chain file is sort of inaccurate there. It could have been a 1bp SNP change (the A->C) and then a 1bp insertion (the ACT->ACTT). I don't know exactly how liftover internals work there, ucsc team may know more
and with the other ones, it looks like https://www.ncbi.nlm.nih.gov/snp/rs2236566#variant_details is outdated and it only exists on an older assembly (NC_000001.9 is chr1 from a older assembly, NC_000001.10 is chr1 on GRCh37) and then the third one rs12728058 is in a very chaotic region of the genome (subtelomere)...this is a zoomed out view of that. your SNP is in the middle of there, so even though NCBI confidently lists where it is on GRCh38, it is maybe difficult for liftover to do so (NCBI uses flanking sequences to help https://www.ncbi.nlm.nih.gov/snp/rs12728058)
Did you see this note on the UCSC liftOver start page? "We do not recommend liftOver for SNPs that have rsIDs. See our FAQ for more information." - Did you have a look at the FAQ?
random note: one could argue that the chain file is sort of inaccurate there. It could have been a 1bp SNP change (the A->C) and then a 1bp insertion (the ACT->ACTT). I don't know exactly how liftover internals work there, ucsc team may know more
and with the other ones, it looks like https://www.ncbi.nlm.nih.gov/snp/rs2236566#variant_details is outdated and it only exists on an older assembly (NC_000001.9 is chr1 from a older assembly, NC_000001.10 is chr1 on GRCh37) and then the third one rs12728058 is in a very chaotic region of the genome (subtelomere)...this is a zoomed out view of that. your SNP is in the middle of there, so even though NCBI confidently lists where it is on GRCh38, it is maybe difficult for liftover to do so (NCBI uses flanking sequences to help https://www.ncbi.nlm.nih.gov/snp/rs12728058)
https://jbrowse.org/code/jb2/v2.12.0/?config=test_data%2Fconfig_demo.json&session=share-nCBvHC3odD&password=ib20C