Notes on the 'Nucleotide-by-nucleotide' plot:

This is a nucleotide-by-nucleotide plot of the likelihood ratio statistic for each reference - non-reference sequence pair and also summed over the phylogenetic tree. Gaps, and stop and start codons for each sequence, are also annotated on the plot. The six panels show the following information:

1. This panel displays the raw likelihood ratio scores at each position in the alignment. There is a seperate track for each reference - non-reference sequence pair (labelled at the right). Gaps, and stop codons in each of the null and alternate model annotated CDSs, for each sequence, are marked on the appropriate tracks.
   
   
2. This panel displays the raw likelihood ratio scores, summed over the phylogenetic tree (i.e. the input list of sequence pairs; details), at each column position in the input alignment.
   
   
3. This panel displays the running mean (i.e. sliding window mean) likelihood ratio scores at each position in the alignment, for each reference - non-reference sequence pair. The sliding window width is annotated on the plot, and can be changed via the 'Redraw plot' link.
   
   
4. This panel displays the running mean of the phylogenetically summed likelihood ratio scores (i.e. running mean of panel 2).
   
   
5. This panel shows the input Query, or alternate model, CDS(s) (red bars) and the input Known, or null model, CDS(s) (blue bars).
   
   
6. This panel shows the phylogenetic sum of sequence divergences (mean number of mutations per nucleotide) for the sequence pairs that contribute to the likelihood ratio sum at each position in the alignment. In any particular column, some sequences may be omitted from the likelihood ratio calculations due to gaps or stop to non-stop transitions. Statistics in regions with lower summed divergence (i.e. partially gapped regions) have a lower signal-to-noise ratio.
   
   

Notes:

• Statistics are only shown within the query region.
• For panel 4, any columns with gaps (or ambiguous nt codes), in any sequence in the input list of sequence pairs, are omitted. Such columns are omitted before taking the running mean. Thus where gaps occur a, for example, 21 nt window includes a total of 21 columns taken from either side of the gap but none from within the gap. On the 'Redraw plots' page, you may choose to extend the plot into partially gapped regions, provided the summed divergence of the contributing sequence pairs in the region is greater than some user-defined threshold value (details).
• In general, you shouldn't have any null or alternate model stops on the reference sequence track. If there are any, then this means that the 'Known CDSs' or 'Query CDSs' (respectively) that you inputted are not in fact ORFs. I.e. you have probably misannotated your CDSs.
• In general, you wouldn't expect to have null model stop codons in any of the sequences. The null model represents known CDSs, which are typically conserved across an alignment. Stops in some sequences near the ends of CDSs are not too unusual and indicate that the CDS terminates early in some sequences. However, if some sequences have many null model stops, then this may indicate a CDS annotation problem.
• If there are no alternate model stops in the non-reference sequences, or any stops are located close to the 'Query CDS' end-point, then this indicates that the 'Query CDS' is conserved as an ORF across the alignment. For a long ORF, this in itself may be strong evidence that the ORF is a CDS. On the other hand, if there are many alternate model stops, then this is evidence against the ORF being a CDS, at least in some sequences.
• Note that alignment problems may cause a non-reference sequence codon to be aligned out-of-frame to a reference sequence codon. This may occassionally result in an out-of-frame non-reference sequence stop codon being incorrectly annotated on the plot. This can be avoided by keeping gaps in groups of three within the null and alternate model CDSs (see also this note). Therefore, if you have an isolated stop codon in what otherwise appears to be a long conserved ORF, you should check that it is not the result of a local alignment problem.
• Note also that in places where the reference sequence contains alignment gaps, there is no frame information for the non-reference sequences. As far as calculation of statistics is concerned, all such regions are omitted. However for the stop and start codon annotation, any non-reference sequence stops or starts within reference sequence gaps will be missed.