HLAProfiler utilizes k-mer profiles to improve HLA calling accuracy for rare and common alleles in RNA-seq data

Table 1 Two-field accuracy and exact sequence matching for novel and partial alleles

	HLAProfiler	OptiType	seq2hla	HLAForest	HLAminer	PHLAT
Novel alleles*
Sequence identified	63%	-	-	-	-	-
Sequence identified or two-field accuracy	68%	-	-	-	-	-
One-field accuracy	97%	15%	17%	19%	37%	96%
Two-field accuracy	25%	4%	4%	3%	10%	26%
Partial alleles
Sequence identified	67%	-	-	-	-	-
Sequence identified or two-field accuracy	85%	-	-	-	-	-
One-field accuracy	97%	95%	94%	19%	72%	97%
Two-field accuracy	85%	41%	21%	3%	23%	46%

Results are based on two sets of simulated data for 100 samples, each having exactly one partial allele or one novel allele. Accuracy for novel alleles is defined as identification of the exact sequence, one-field accuracy, or two-field accuracy
*In the case of novel alleles, the correct protein sequence can be identified without correctly identifying one- or two-field precision, or one- or two-field precision can be identified while missing the exact protein sequence

ISSN: 1756-994X