Comparison of assembly algorithms for improving rate of metatranscriptomic functional annotation

Microbiome

Table 1 Overlap in assemblies

	Metavelvet, k= 39	Metavelvet, k= 51	Oases, k= 27–35	Oases, k= 39–45	Oases, k= 51–53	Trinity default
All single-end contigs
Metavelvet, k = 27	85.50%	69.60%	76.80%	76.20%	59.70%	98.00%
Metavelvet, k = 39		84.70%	64.50%	65.00%	71.20%	96.80%
Metavelvet, k = 51			64.30%	68.00%	48.10%	96.60%
Oases, k = 27–35				85.60%	68.10%	95.40%
Oases, k = 39–45					91.00%	97.30%
Oases, k = 51–53						97.60%
BLAST score >50
Metavelvet, k = 27	87.20%	79.60%	72.80%	78.40%	74.50%	97.00%
Metavelvet, k = 39		89.80%	65.50%	64.70%	82.30%	96.50%
Metavelvet, k = 51			72.90%	69.50%	48.20%	97.10%
Oases, k = 27–35				92.60%	85.70%	88.00%
Oases, k = 39–45					92.00%	92.70%
Oases, k = 51–53						92.60%

Single-end assemblies constructed from 516,881 single-end reads of putative bacterial mRNA origin obtained from a non-obese diabetic (NOD) mouse cecal sample were evaluated on the uniqueness of the reads incorporated into contigs. Figures indicate the percentage of reads of the smaller dataset that are incorporated into contigs in both datasets.

ISSN: 2049-2618