Pre–messenger RNA nucleotide replaced by the polyadenylation [poly(A)] tail in different species^a.

a<p>Each of the upstream 100–base messenger RNA (mRNA) sequences directly adjacent to the poly(A) tail starting position is unique, but all the mapped genomic sites (1.36 sites on average per unique animal or plant mRNA) from a single unique mRNA were counted.</p>b<p>The observed percentage of pre-mRNA adenosine replaced by the poly(A) tail is the genomic adenosine frequency at the site corresponding to the first adenosine of mRNA poly(A) tails. The statistical significance marked on the poly(A) site adenosine frequency was based on the chi-square test using the observed values (number of observed A and number of observed non-A) against the random model theoretical values (number of theoretical A and number of theoretical non-A).</p>**<p>significance at <i>P</i><0.01 in <i>N. crassa</i>, at <i>P</i><0.001 in <i>P. falciparum</i> and <i>S. pombe</i>, and at <i>P</i><0.000,000,0001 in all other species.</p>c<p>The average A nucleotide content in the 100 nucleotides upstream of the poly(A) site of mRNA. The poly(A) site percentage is not correlated with the mRNA adenosine content (<i>r</i><0.09).</p>d<p>The theoretical adenosine poly(A) site frequency in the alignment from the random model. If the A nucleotide percentage in mRNA is <i>p</i>, the adenosine poly(A) site from the alignment will be <i>p</i>+<i>p</i>(1−<i>p</i>) = <i>p</i>(2−<i>p</i>), where (1−<i>p</i>) is the non-A nucleotide content (See <b><a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0079511#pone.0079511.s004" target="_blank">File S1</a></b>). The observed poly(A) site adenosine frequency is clearly not random; it is significantly higher (<i>P</i> = 0) than its random model value.</p><p>Note that the nucleotide at the mRNA poly(A) site is usually an A and occasionally a U. Poly(A) sites with C or G are rare.