T3/Wheat Practical Haplotype Graph (PHG) imputation

PHG Description

The Practical Haplotype Graph (PHG) database was constructed using whole-exome capture sequencing data from a diverse set of wheat accessions.

Imputed protocols and projects
Method: Practical Haplotype Graph - PHG V2 with 2.9M markers
Visualize: Explore options
Description: Imputation pipeline and accuracy

Protocol	Download genotype project	accessions	Cluster with PHG	Imputation accuracy vs 2019_HapMap
Infinium 90K v2.1	TCAP90K_HWWAMP	295	PCA	92% - details
	TCAP90K_SpringAM_panel	245
	TCAP90K_YQV14	216
	TCAP90K_CSRVAL14	116	PCA
	VRND4_UCD_2015	76	PCA
	TCAP90K_NAMparents_panel	56
	TCAP90K_HWWAMP_SRPN	16
	TCAP90K_SWWpanel	313
	TCAP90K_LeafRustPanel	335
	TCAP90K_SNBWWW	266	PCA
Infinium 9K v2.1	NSGCwheat9K_4X	542	PCA	94% - details
	WorldwideDiversityPanel_9K	2255
GMS	USU_2021	288		82% - details
	USU_2022	471
MRASeq RPN	RPN_2020	86	PCA	84% - details
	RPN_2021	90	PCA
	RPN_2022	80	PCA
	RPN_2023	88	PCA
	RPN_2020-23	286	PCA
	RPN_2020-21	157	PCA
	RPN_2021-22	150	PCA
	RPN_2022-23	151	PCA
MRASeq RGON	RGON_2021	300	PCA	84% - details
	RGON_2022	327	PCA
	RGON_2023	412	PCA
UCD GBS	UCD_2022	345	PCA	na
SDSU GBS	SDSU_2022	157	PCA	na
	SDSU_2023	155	PCA
MSU GBS	MSU_2023	3115	PCA	na
ThermoFisher AgriSeq	UCD_2023_AgriSeq_1	92	PCA	na
	UCD_2023_AgriSeq_2	94	PCA
	WSU_2023_AgriSeq_A	96		too many ambiguous alt alleles
	WSU_2023_AgriSeq_B	96
USDA 3K	MNA_2022	765		93% - details
	MonSU_2022	417
	NDS_2022	124
	UCD_2022	176
	MonSU_2023	432
	UWM_2023	181
	WSU_2023	364
Allegro UCD	UCD2020_Allegro_CAD	179

Notes:
1. Cluster analysis for Infinium calculated by merging unimputed data with PHG genotypes (common markers).
2. Cluster analysis for other protocols (no common markers) calculated on the unimputed data and showing which accessions are in the PHG
3. Imputation accuracy calculated by comparing every 30th marker between imputed output and 2019_HapMap for matching accessions

Imputation Pipeline

The imputation was done using the ImputePipelinePlugin of the PHG (PHG Wiki). The ImputePipelinePlugin was run with minRead=0 so that all haplotypes are imputed. PHG Version 0.0.40 was used to run the ImputePipelinePlugin.

Imputation accuracy

Genotyping Protocol	PHG accession	other accessions
Genotyping by multiplexed sequencing, GMS
Infinium 9K	93%	71%
Infinium 90K	94%	71%

PHG imputation protocol

Converting Illumina Infinium before imputation - Illumina data is saved in the T3 database as Ref = A_allele, Alt = B_allele. To impute this data and merge with other genotype protocols it is necessary to convert the Illumina so that it aligns (strand and orientation) with the reference genome. Detailed description

Imputation documentation - Testing PHG accuracy

GitHub repository of scripts to prepare and test data - PHG v2