Global, highly specific and fast filtering of alignment seeds

BMC Bioinformatics

Table 2 Runtime and memory requirements of the methods when run with best weight as determined by Fig. 4

Method	Weight	Patterns	Sensitivity	\(\widehat{\text {FP}}\)	Runtime (min)	Memory (GB)
M1	13	1	0.909	34.3	32	38
M2	15	1	0.914	2.99	17	26
M3	17	2	0.906	0.375	38	64
M3	18	4	0.909	0.0116	87	156
M4	12	4	0.928	0.349	194	80
M5	15	4	0.955	0	75	137

The weights were chosen such that each method had the lowest possible \(\widehat{\text {FP}}\) but a sensitivity of at least 0.9. Note that the runtime and memory for M4 cannot be directly compared to the other methods as we needed to split the data into batches and do multiple sequential runs because of memory limitations

ISSN: 1471-2105