biobambam: tools for read pair collation based algorithms on BAM files

Table 4 Run-time comparison of bamtofastq and alternatives on compute farm nodes (part b)

Run-time comparison for BAM to FastQ conversion on server blades
Data set	Program	Memory/GB	Run-time/minutes
ERR328876	biobambam	0.13	26.90±0.49
	bam2fastq	0.85	74.55±0.60
	bampe2fqworphans	0.70	51.10±0.61
	Picard	4.95	137.98±1.58
	bamUtil		≥1440
ERR054938	biobambam	0.13	29.42±0.57
	bam2fastq	0.98	76.15±1.54
	bampe2fqworphans	0.84	50.8±0.62
	Picard	6.69	152.62±1.06
	bamUtil	6.13	440.8±5.68
ERR328190	biobambam	0.13	51.70±0.60
	bam2fastq	4.21	106.73±0.45
	bampe2fqworphans	3.45	74.71±0.87
	Picard	16.12	170.88±1.07
	Picard_1,16	16.12	123.10±8.04
	Picard_1,230	30.94	120.06±8.64
	bamUtil		≥1440
SRP017681	biobambam₁₈	0.18	137.21±1.17
	biobambam₂₃	2.45	61.49±0.68
	bam2fastq	6.56	153.85±1.74
	bampe2fqworphans	5.20	91.83±0.49
	Picard	16.10	261.24±3.05
	Picard_1,16	16.10	200.63±2.61
	Picard_1,230	28.45	200.60±2.74
	bamUtil		≥1440
ERP001231	biobambam	0.13	111.52±0.77
	bam2fastq	4.35	349.38±1.18
	bampe2fqworphans	3.48	229.80±1.30
	Picard	14.81	489.11±4.40
	bamUtil		≥1440

Run-time comparison of biobambam’s bamtofastq, bam2fastq, bampe2fqworphans, Picard’s SamToFastQ and bamUtil’s bam2fastq for the data sets ERR328876, ERR054938, ERR328190, SRP017681 and ERP001231 described in Table2 on compute farm nodes. For the data set SRP017681 bamtofastq was run with a default hash table size of 2¹⁸ and an increased size of 2²³ for comparison. bamUtil dedup failed to process the data sets ERR328876, ERR328190, SRP017681 and ERP001231 within the 24 hour limit. Picard used close to 16 GB of memory for the data sets ERR328190 and SRP017681. We have verified that no significant speed ups can be obtained by allowing more memory in a reduced concurrency setting, where we have run only a single process at a time using 16 or 230 GB of memory on an otherwise idle machine. For both data sets the maximum amount of memory used by Picard when provided with 230 GB was significantly less than what was available.

ISSN: 1751-0473