Predicting bacterial resistance from whole-genome sequences using k-mers and stability selection

BMC Bioinformatics

Table 1 M. tuberculosis study: cross-validation results

	R / S	k-mers	AUC	p
ethambutol	333 / 712	8 (79)	91.0 (92.0)	0.75
ethioniamide	172 / 250	3 (9)	82.2 (85.3)	0.7
isoniazid	815 / 472	4 (5)	96.2 (96.0)	0.8
kanamycin	187 / 484	2 (49)	90.1 (93.2)	0.8
ofloxacin	238 / 458	1 (1)	91.2 (91.1)	0.8
rifampicin	668 / 533	7 (6)	96.6 (96.6)	0.8
streptomycin	492 / 678	7 (22)	90.8 (92.3)	0.8

R / S: number of resistant and susceptible strains. k-mers: number of k-mers in final models obtained with the stability selection approach and with the classical lasso penalty (between brackets). AUC: cross-validated AUC obtained with the stability selection approach and with the classical lasso penalty (within brackets). p: probability threshold selected for the stability approach

ISSN: 1471-2105