Handle differences in N-padding

[charlesgregory]

Ensembl adds N-padding to some of the PATCH contigs to make their coordinates during alignment identical to the position on the primary contig the patch is derived from:

All alternative assembly and patch regions have their sequence padded 
with N's to ensure alignment programs can report the correct index
regions

e.g. A patch region with a start position of 1,000,001 will have 1e6 N's added
its start so an alignment program will report coordinates with respect to the
whole chromosome.

Example contig from the HG-2365 PATCH:

Reference Build	Name	Length
Ensembl	CHR_HG2365_PATCH	102714182
gencode	ML143371.1	5500449
NCBI	NW_021160017.1	5500449

So for this contig Ensembl has added 97,213,733 extra Ns to the sequence. Which means we cannot match it with gencode's ML143371.1 or NCBI's NW_021160017.1.

[charlesgregory]

In order to address this we would have to amend the file format of the recontig mapping files to have an optional column that contains the offset to be applied to record starting positions.