Table 4.

Mutation spectrum at centromeres is unlike that of gene conversion

Study	Ts: Tv	Ts:Tv freq	1 bp deletions	Del:BPS freq	1 bp insertions	Other
CDEII	46: 68	0.68	4	0.047	0	18a
Genome-wide (WT)	12: 19	0.63	1	0.032	0	1
URA3 Lang (WT)	46: 121 *	0.38	22	0.017	3	15
CAN1 Lang (WT)	65: 85	0.76	56	0.034	8	13
URA3 Hicks (WT)	19: 27	0.70	5	0.014	1	9
URA3 Hicks (GC)	32: 24 *	1.33	32	0.074	1	14

The mutation spectrum of CDEII is compared to the wild type spontaneous mutation (WT) inferred from various studies (Genome-wide, [15]; Lang, [23]; Hicks, [13]), and the mutation spectrum expected under gene conversion (GC, [13]). "Other" indicates indels that are -> 1 bp long, and in the case of CDEII also includes indels occurring in homopolymer runs that were 5 bp or longer (N = 14), other mutations described as "Other" for the WT and GC datasets are described in [13]. The deletion: base pair subsitution frequency (Del:BPS) corrects for the difference in target length at which deletions could be observed compared to the target length for base pair substitutions ([23], e.g. for URA3, indel target size = 804 bp, and BPS target size = 104 bp). For CDEII the number of deletions (N = 4) was compared to the number of base pair substitutions (N = 84) outside homopolymer runs. * Significantly different from CDEII (Fisher's exact test, P < 0.05). Poisson tests showed that none of the Del:BPS frequencies were significantly different from that of CDEII.