GT-WGS: an efficient and economic tool for large-scale WGS analyses based on the AWS cloud service

BMC Genomics

Table 1 Comparison with previous benchmarks of time and cost for WGS data analysis based on different pipeline and hardware

Tool	Aligner + Variant Caller	Depth	Time	Cost	Depth^a	Time^a	Cost^a	Hardware
Genomekey + COSMOS [8]	BWA + GATK HaplotypeCaller	37×	4.9 h	$48.5	55×	7.3 h	$72.1	20× AWS c2.8xlarge
Churchill [8]	BWA + GATK UnifiedGenotyper	30×	1.7 h	–	55×	3.1 h	–	16× AWS r3.8×large
STORMseq [8]	BWA + GATK lite	38×	176 h	$32.8	55×	255 h	$47.5	–
Crossbow [9]	Bowtie + SOAPsnp	38×	4.5 h	$71.4	55×	6.5 h	$103.3	20× AWS c1.xlarge
Crossbow [9]	Bowtie + SOAPsnp	38×	2.5 h	$83.6	55×	3.6 h	$121	40× AWS c1.xlarge
PEMapper / PECaller [10]	PEMapper + PECaller	30×	29.3 h	–	55×	53.7 h	–	–
Globus [11]	Bowtie2 + GATK	30×	12 h	–	55×	22 h	–	1× AWS cr1.8xlarge
SevenBridges [12]	BWA + GATK	15×	8 h	$14.1	55×	29.3 h	$51.7	–
BGI-online (BALSA) [13]	BALSA	50×	5.5 h	–	55×	6 h	–	6-core CPU, 64GB RAM, GPU GTX680
Average						42.9 h	$79.1

‘^a’ means time and cost of different depth data are normalized to 55× with linear relationship. ‘-’ means not reported

ISSN: 1471-2164