A-GAME: improving the assembly of pooled functional metagenomics sequence data

BMC Genomics

Table 2 Comparison of workflows for the assembly and annotation of eDNA insert data using Lam et al. [16] pooled inserts

	Insert assignment based on end-tags				Completeness			Computational requirements
	Complete^a	2 of 2 ends^b	1 of 1 ends^c	Partial^d	% assembled^e	% of reference proteins^f	Assembly N50	CPU time (h)	RAM peak (Gb)
original assembly	40	6	10	11	100.00	100.00	36,113	NA	NA
SPAdes (F1)	34	13	9	11	88.16	86.35	34,329	2.03	5.3
Velvet (F2)	18	27	10	12	66.58	64.79	32,942	1.67	6.61
MEGAHIT (F3)	30	19	8	10	95.14	92.13	34,446	1.21	3.22
MetaVelvet (F4)	19	26	9	13	74.64	73.38	33,150	1.75	7.01
meta-SPAdes (F5)	34	13	9	11	88.16	86.35	34,329	2.03	5.3
MOCAT2	19	27	8	13	67.62	65.92	25,246	1.91	4.51
Parallel META2	12	34	6	15	40.48	37.87	26,408	2.36	3.11
Original from LAM et al	19	28	7	13	72.47	69.97	33,347	NA	NA

^aInsert assembled into a single contig matching both end tags
^bInserts assembled into multiple contigs, both end tags are assigned
^cInserts for which only a single end tag is available and gets assigned
^dInserts for which both ends are available but only one is assigned
^ePercentage of reference assembly represented in the pooled assembly
^fPercentage of proteins from the reference assembly recovered in the pooled assembly

ISSN: 1471-2164