I have downloaded cram files, but I don't know which exact version of hg38 was used to align the reads. Can you find the corresponding fasta if you have the M(D)5 strings? For now, it seems I can only test and see when it does. Of course the obvious solution is to ask the person who generated it, but that does not always work out. A part of the header looks like:
@SQ SN:chr1 LN:248956422 M5:6aef897c3d6ff0c78aff06ac189178dd UR:/scratch/hg38.fa
@SQ SN:chr2 LN:242193529 M5:f98db672eb0993dcfdabafe2a882905c UR:/scratch/hg38.fa
@SQ SN:chr3 LN:198295559 M5:76635a41ea913a405ded820447d067b0 UR:/scratch/hg38.fa
@SQ SN:chr4 LN:190214555 M5:3210fecf1eb92d5489da4346b3fddc6e UR:/scratch/hg38.fa
@SQ SN:chr5 LN:181538259 M5:a811b3dc9fe66af729dc0dddf7fa4f13 UR:/scratch/hg38.fa
@SQ SN:chr6 LN:170805979 M5:5691468a67c7e7a7b5f2a3a683792c29 UR:/scratch/hg38.fa
@SQ SN:chr7 LN:159345973 M5:cc044cc2256a1141212660fb07b6171e UR:/scratch/hg38.fa
@SQ SN:chr8 LN:145138636 M5:c67955b5f7815a9a1edfaa15893d3616 UR:/scratch/hg38.fa
@SQ SN:chr9 LN:138394717 M5:1b79085d423b806957b7564497cac5e4 UR:/scratch/hg38.fa
@SQ SN:chr10 LN:133797422 M5:c0eeee7acfdaf31b770a509bdaa6e51a UR:/scratch/hg38.fa
@SQ SN:chr11 LN:135086622 M5:1511375dc2dd1b633af8cf439ae90cec UR:/scratch/hg38.fa
@SQ SN:chr12 LN:133275309 M5:96e414eace405d8c27a6d35ba19df56f UR:/scratch/hg38.fa
@SQ SN:chr13 LN:114364328 M5:787e7eb2d9187bbc20334062332569d4 UR:/scratch/hg38.fa
@SQ SN:chr14 LN:107043718 M5:e0f0eecc3bcab6178c62b6211565c807 UR:/scratch/hg38.fa
@SQ SN:chr15 LN:101991189 M5:f036bd11158407596ca6bf3581454706 UR:/scratch/hg38.fa
@SQ SN:chr16 LN:90338345 M5:9adbaf8ef0094c71470e87eb18e9b5d4 UR:/scratch/hg38.fa
@SQ SN:chr17 LN:83257441 M5:f9a0fb01553adb183568e3eb9d8626db UR:/scratch/hg38.fa
@SQ SN:chr18 LN:80373285 M5:11eeaa801f6b0e2e36a1138616b8ee9a UR:/scratch/hg38.fa
"sometimes", the sequences are hosted at the EBI. For example your first sequence with md5 checksum = 6aef897c3d6ff0c78aff06ac189178dd is available (not fasta but plain string) at:
https://www.ebi.ac.uk/ena/cram/md5/6aef897c3d6ff0c78aff06ac189178dd
see REF_PATH and REF_CACHE in the samtools manual.
Aha good start, but doesn't work for each of the chromosomes. Seems the fasta is "special", then. I found someone on Twitter to point me to the right one (https://ftp.1000genomes.ebi.ac.uk/vol1/ftp/data_collections/1KG_ONT_VIENNA/reference/1KG_ONT_VIENNA_hg38.fa.gz)
Log in to answer this question.
Googling for the checksum value leads to ENA Browser pages that also have the MD5 sums on the page for the relevant chromosomes (for example):
https://www.ebi.ac.uk/ena/browser/view/CM000664
https://www.ebi.ac.uk/ena/browser/view/CM000679
Great, that led me to GCA_000001405... However, it is not correct for all chromosomes. For example, that chromosome 13 (https://www.ebi.ac.uk/ena/browser/view/CM000675.2) has an MD5 checksum of a5437debe2ef9c9ef8f3ea2874ae1d82, while the cram I have has 787e7eb2d9187bbc20334062332569d4 :-(
I found someone on Twitter to point me to the right one (https://ftp.1000genomes.ebi.ac.uk/vol1/ftp/data_collections/1KG_ONT_VIENNA/reference/1KG_ONT_VIENNA_hg38.fa.gz).
Not sure if there could be a better way :)