Karyotypically Sorted Genome Assembly

Question

[blackgore]

How does GATK define their requirement of ‘karyotypic’ sorting? In the simple case of the UCSC human assembly, yes it is to place the chromosomes in numerical order 1-22, followed by X,Y and M, but where does one place the additional unlocalised contigs that are present in, for example, the Ensembl assembly? This was brought up as a comment in a related post.

Answer 1

[Brad Chapman]

The Broad bundles, mentioned in your referenced post, have them pre-sorted for you and include all of the unlocalized and unplaced contigs. If you are building your own you can download the index and directly examine the ordering:

$ wget ftp://[e-mail]/bundle/1.2/hg19/ucsc.hg19.fasta.fai.gz
$ gunzip ucsc.hg19.fasta.fai.gz
$ cat ucsc.hg19.fasta.fai | cut -f 1
[...]
chr22
chrX
chrY
chr1_gl000191_random
chr1_gl000192_random
chr4_ctg9_hap1
chr4_gl000193_random
chr4_gl000194_random
chr6_apd_hap1
chr6_cox_hap2
chr6_dbb_hap3
[...]

[Brad Chapman]

Sorry, I missed that you were interested in the GRCh37/Ensembl naming. Broad also has a bundle for that; it's definitely worth digging on their FTP site before creating your own: ftp://[e-mail]/bundle/1.2/b37/human_g1k_v37.fasta.fai.gz

[blackgore]

Thanks Brad for the quick response. I've done as you suggested, and mapped the ordering of the files from Broad to the Ensembl release. There is a greater number of contigs listed in the Ensembl file than the .fai file mentioned above, but it's a great place to start.

[blackgore]

Thanks Brad, I'd not realised that Broad supplied sets for the Ensembl release too. Thanks for your advice!

Answer 2

[Cyriac Kandoth]

The reference FASTAs downloadable from Ensembl are not karyotypically sorted, but that can be fixed with a little bit of Perl...

Download and unzip the Build37 (hg19) reference FASTA file from Ensembl release 72:

curl -LO ftp://ftp.ensembl.org/pub/release-72/fasta/homo_sapiens/dna/Homo_sapiens.GRCh37.72.dna.primary_assembly.fa.gz

gunzip Homo_sapiens.GRCh37.72.dna.primary_assembly.fa.gz

Use this Perl one-liner that splits the FASTA file per chrom/contig into a temporary folder, and then concatenates them in karyotypic order:

perl -e 'use File::Temp qw/tempdir/; use IO::File; $d=tempdir; $fh; map{if(m/^\>(\S+)\s/){$fh=IO::File->new(">$d/$1.fa");} print $fh $_;}`cat Homo_sapiens.GRCh37.72.dna.primary_assembly.fa`; foreach $c(1..22,X,Y,MT){print `cat $d/$c.fa`}; print `cat $d/GL*`' > Homo_sapiens.GRCh37.72.dna.primary_assembly.reordered.fa