Table 3. A description of the top three k-mers found by AdaBoost and their corresponding regions in A. baumannii AB_2008-15-34-7, S. aureus 08-01059, S. pneumoniae ATCC 700669, and SMRU2064.

Rank	α-value	k-mers with an identical pattern	Corresponding genes	PATRIC annotation
A. baumannii, carbapenem
1	1.21	1	fig|1221255.3.peg.3516	LysR-family transcriptional regulator clustered with PA0057*
2	0.82	4	fig|1221255.3.peg.3314	NAD+–asparagine ADP-ribosyltransferase
3	0.77	3	fig|1221255.3.peg.631	Dihydrodipicolinate synthase family
S. aureus, methicillin
1	2.37	3321	fig|1413344.3.peg.2510,fig|1413344.3.peg.2511,fig|1413344.3.peg.2512	Glycerophosphoryl diester phosphodiesterase (EC 3.1.4.46); MaoC domain protein; Penicillin-binding protein PBP2a, Penicillin-binding protein PBP2a, methicillin resistance determinant MecA, transpeptidase
2	1.16	26	fig|1413344.3.peg.1752	hypothetical protein
3	0.84	17	fig|1413344.3.peg.1698	Phage protein
S. pneumoniae, beta-lactam
1	0.74	17	fig|561276.4.peg.338	Cell division protein FtsI [Peptidoglycan synthetase] (EC 2.4.1.129)
2	0.62	16	intergenic region	between Multiple sugar ABC transporter proteins (fig|561276.4.peg.108 and fig|561276.4.peg.109)
3	0.60	14	fig|561276.4.peg.338	Cell division protein FtsI [Peptidoglycan synthetase] (EC 2.4.1.129)
S. pneumoniae, co-trimoxazole
1	0.66	5	intergenic region	immediately downstream of Dihydropteroate synthase (EC 2.5.1.15) (fig|1313.2194.peg.17)
2	0.55	2	fig|1313.2194.peg.1876	Dihydrofolate reductase (EC 1.5.1.3)
3	0.51	6	fig|1313.2194.peg.1874	Glucan-binding domain / Lyzozyme M1 (1,4-beta-N-acetylmuramidase) (EC 3.2.1.17)

Genomes were chosen as examples with exact k-mer matches. The complete list of k-mers is described in the supplementary data file.

^*Occurs next to fig|1221255.3.peg.3517, Metallo-beta-lactamase superfamily protein PA0057.