how to add non-reference genes to BED file of intervals for PHG

Question

[b.contreras.moreira]

Hi, one of the reasons to build a pangenome is to capture genes which might be absent in the reference genome of interest. While following the documentation of the PHG you get to a point where you need to "Create bedfile of intervals for PHG reference ranges".

My question is how to add non-reference genes to such a file, should I do it by adding genomic ranges in the reference genome that will roughly match the equivalent regions in other assemblies that contain those non-reference genes?

Consider the following region from chr3H in barley adapted from barley_pangenes:

chr	start	end	reference gene	assembly2 gene
3H	331491	334552	HORVU.MOREX.r3.3HG0218270	Horvu_10350_3H01G001500
3H	NA	NA	NA	Horvu_10350_3H01G001600
3H	414358	417904	HORVU.MOREX.r3.3HG0218320	Horvu_10350_3H01G001700

• Should I merge these gene models and create a custom range such as 3H:331491-417904 to make sure reads matching non-reference gene Horvu_10350_3H01G001600 are used to build haplotypes?

Thanks for your help, Bruno

Answer 1

[lcj34]

hi Bruno - You should not need to create custom ranges. The anchorwave alignment process aligns full genomes to each other using collinear regions identified through GFF file CDS and exon regions. AnchorWave has the ability to identify novel anchors within long inter-anchor regions. Based on this we expect all, or nearly all, of each assembly genome's sequence to be included as haplotype sequence aligned to either a reference defined conserved or non-conserved region. Later, when imputing against the pangenome that includes the haploypes from all assemblies, you should be able to align to these regions.

You can read more about anchorwave from the readme at the gihub page (https://github.com/baoxingsong/AnchorWave) or from the Anchorwave paper (https://www.pnas.org/doi/abs/10.1073/pnas.2113075119)