Fulgor: a fast and compact k-mer index for large-scale matching and color queries

Table 8 Quality of pseudoalignment algorithms querying 100,000 simulated reads against 50,000 Salmonella genomes indexed with Fulgor

% Positive	Full-intersection		Threshold-union		Kallisto		Alevin-fry
% Positive	TPR	FPR	TPR	FPR	TPR	FPR	TPR	FPR
90	95.0	27.0	97.7	30.0	95.0	27.0	95.1	27.0
70	95.1	27.0	97.7	30.0	95.1	27.0	95.1	27.0
25	95.1	27.0	97.7	30.0	95.2	27.0	95.2	27.0
10	95.5	27.0	97.8	30.0	95.5	27.0	95.5	27.0

We vary the percentage of positive reads simulated from indexed Salmonella genomes by diluting queried read sets with negative reads simulated from a reference human transcriptome. We consider a mapped positive read (deriving from indexed references) to be a true positive if the reference of origin is in the returned set of compatible references; and a mapped negative read (deriving from human chromosome 19) to be a false positive. We denote true and false positive rates (%) to be TPR and FPR, respectively. For the threshold-union method, we use \(\tau =0.8\)

ISSN: 1748-7188