Assessing optimal: inequalities in codon optimization algorithms

Table 1 Descriptions of codon usage databases for either the generic class or B strain of E. coli. Each annotation describes the source of the genetic data, the total number of coding DNA sequences (CDS) extracted from the gene source(s), and the number of codons extracted from genes used to construct each database

Author or database	*E. coli* strain	Gene source	# CDS	# codons
Sharp and Li [8] ^a	Generic	GenBank	27	6240
			15	9223
			57	25,010
			58	22,612
Kazusa database [9]	Generic	GenBank	8087	2,330,943
Kazusa database [9]	B	GenBank	11	3771
HIVE-CUT database [10]	Generic	GenBank and RefSeq	68,262,063	20,219,118,236
HIVE-CUT database [10]	B	GenBank and RefSeq	13,042	3,953,593
GtRNAdb [11, 12]	Generic	GenBank and RefSeq	5011	1,538,003
GenScript^b	Proprietary	Undefined	Undefined	Undefined
Dong et al. [13]	W1485 (K12)	N/A	Total RNA	Undefined

^aAuthors divided their dataset into four groups that represent genes that exhibit “very high expression,” “high levels of expression,” “moderate codon bias,” or “low codon bias” that are represented from the top down, respectively
^bwww.genscript.com

ISSN: 1741-7007