| This Article | |
 | |
 |
|
| |
| Services | |
| |
|
|
|
|
| ||||||||||||||||
|
Borevitz et al. 10.1073/pnas.0705323104. |
Fig. 5. Thresholds used for SFP calls reveal an excess of SFPs with significantly reduced hybridization intensities compared with the Col reference genome. Duplication SFPs do not have a known physical position and are treated as missing data (NA), as are SFPs with statistics that cannot be rejected as different from Col but are also not likely to be the same as Col (NA). Tests without clear hybridization polymorphism are called the same as Col, nonSFP. SFP allele frequency across up to 23 accessions. SFP allele frequencies are skewed toward rare variants under this calling scheme, because each accession must be detected as different from Col. A mirror image is expected where many rare Col alleles would be identified as SFPs across all accessions, giving a U-shaped distribution. We are limited, however, by the design of the array and the current statistical model.
Fig. 6. Multi-SFPs. Alleles are rare as determined with Sequence Data 306/373 (82%) of SFPs that contained SNPs are biallelic as determined across 16 accessions and up to 1,563 features, for which SFP calls and dideoxy sequence data were available. Empirical observations of multi-SFP alleles is higher than expected by chance but was modeled (1) and reflects the fact that high-nucleotide diversity regions (>1%) exist at a low level in Arabidopsis.
1. Jiang R, Marjoram P, Borevitz JO, Tavare S (2006) Genetics 173:2257-2267.
Fig. 7. Phylogenetic relationship between the 23 Arabidopsis accessions based on SFPs. A correlation clustering dendrogram was generated with SFP genotypes, which measure changes from the reference strain. This distance measure used was 1 minus the correlation. Bootstrap values showing branch strength greater than 60 of 100 bootstrap samples of SFP genotype data are shown. The two deepest clade group by experiment, while significant genome-wide relatedness can be seen among accessions at lower levels of the tree due to population structure.
Fig. 8. Increased diversity and selection on gene families. The vertical black lines demark 95% of the gene position shuffled permutation null distribution, whereas the vertical blue lines represent a <2% FDR. Row 1 shows the empirical distribution of diversity in 50-kb windows vs. the null for Total Diversity (Column 1), Pairwise Diversity (Column 2), and Tajima's D (Column 3). Row 2 shows the distribution of diversity and selection in NBS-LRR R gene, RLPs, and S-locus protein families. Row 3 shows the distribution of diversity and selection in F-box, Ring Finger, and bHLH gene families.
Fig. 9. Recombination rate vs. diversity Overall a positive relationship was seen. Many outliers exist, however at low recombination regions (0 and 1 cM/Mb) where high diversity reflects the high variation in heterochromatic regions (red oval). The distribution of recombination rate variation across gene windows is shown as kb/cM or cM/Mb from ref. 1.
1. Singer T, Fan Y, Chang HS, Zhu T, Hazen SP, Briggs SP (2006) PLoS Genet 2:e144.
Fig. 10. Diversity and Tajima's D along the whole genome. NB-LRR R genes are shown in blue, RLPs are shown in green, and S-locus proteins are shown in purple. Genome is displayed in 1-Mb pages.
Fig. 11. Diversity and Tajima's D identify candidate RLPs. NB-LRR R genes are shown in blue, and RLPs are shown in green. RLPs under high-diversity regions warrant further molecular and functional study.
| ||||||||||||||||
|