Table 5. Generalized linear model of indels.

	Polymorphic human		Fixed human		Chimp
	p-value<10−206		p-value<10−44		p-value<10−91
	Effect	p-value	Effect	p-value	Effect	p-value
Ln(recombination)
AT→GC SNPs	+	2.5×10−5 ***
Othera SNPs	+	9.8×10−15 ***
AT→GC fixed hum			+	8.2×10−10 ***
Othera fixed hum	+	0.0097 **	+	1.0×10−10 ***	+	0.0035 **
AT→GC chimp					+	4.1×10−4 ***
Othera chimp			+	0.079 .	+	<10−15 ***
GC	−	2.2×10−5 ***			−	0.0058 **
polyAT	+	9.6×10−4 ***	+	1.3×10−6 ***	+	0.015*
CpG	+	0.05 *
%CDS	−	<10−15 ***	−	5.9×10−12 ***	−	1.8×10−12 ***
%UTR			+	0.0083 **	+	0.042 *
SINE	−	1.9×10−9 ***	−	0.017 *	−	0.0013 **
LINE
Repeat maskedb
Chromosome X	−	9.1×10−11 ***
Telomere distance	−	0.0074**
Not scanned
ID	−	2.4×10−9 ***			−	1.6×10−4 ***
ID : Ln(recombination)	+	0.096 .
ID : AT→GC SNPs
ID : Othera SNPs
ID : AT→GC fixed hum			−	0.064 .
ID : Othera fixed hum
ID : AT→GC chimp
ID : Othera chimp
ID : GC	+	8.4×10−4 ***			+	2.1×10−4 ***
ID : polyAT	+	7.8×10−4 ***			+	0.0023 **
ID : CpG
ID : %CDS
ID : %UTR
ID : SINE
ID : LINE
ID : Repeat maskedb
ID : Chromosome X
ID : Telomere distance			−	0.027 *
ID : Not scanned

^asingle nucleotide changes not AT→GC.

^bnot due to SINEs or LINEs

Summary of the variables affecting the number of indels in the data. Significance of explanatory variables in generalized linear models for counts of indels are reported. The effect is only reported as decreasing or increasing number of insertions (deletions). An empty box indicates nonsignificance. “ID” indicates whether deletions or insertions are predicted. See text for interpretation of the positive effect of the “ID:GC” and “ID:poly AT” terms.