For each gene, the number of unique insertions sites observed, absolute count of sites in the positive strand, in negative strand and in both strands is recorded.
To test the differences in the number of variable positions in base-paired sequences (structured regions) compared with non-structured sequences, we used a Chi-squared test to compare the count of sites in structured and non-structured sequence classes lacking perfect sequence conservation across all studied species.
Given a multiple sequence alignment data d from 3 taxa at v homologous sites, i.e. d ∈ {Y, R}3 × v, the likelihood function over the tree space T k is simplified from (2) as follows: (5) l d (t k ) = ∏ q = 1 v l d ·, q (t k ) = ∏ i = 0 7 (l i (t k ) ) c i, where l i (k t) is the likelihood of the the i-th site pattern as in (4) and c i is the count of sites with pattern i.
The total count of 17,400 sites excludes 84 sites classified as oil sites: 77 sites containing ≥6 oil wells with initial production before 1990 (not likely to be multi-well pads characteristic of shale development) and 7 newer sites containing 11 68 wells but which appeared on GoogleEarth to be distributed oil wells and not multi-well production sites.
These make the counting of sites and differences challenging.
See methods section for further detail on estimating counts of sites.
Given the parameter values, the probabilities of the five site patterns are calculated using equation (1) and the counts of sites at each locus (n i0, n i1, n i2, n i3, n i4) are generated by sampling from the multinomial distribution.
Under the Jukes and Cantor (1969) mutation model, the sequence alignments at any locus i can be summarized as the counts of sites, D i = { n i 0, n i 1, n i 2, n i 3, n i 4 }, for five site patterns xxx, xxy, yxx, xyx, and xyz, where x, y, and z are any different nucleotides.
In all binding site classes, the count of binding sites in S. pombe is low and will be excluded from further analysis.
In raw count of unique sites with sequencing coverage, predicted sites were the most common for all enzymes except EcoRV in maize and HincII in rice.
But even if PCB's were excluded from the count, thousands of sites would still show risky levels of other chemicals, like mercury, metals, and pesticide residues, the report said.
