Full Site Detail

Each site in the list above is also explored in detail in the output. The highest scoring site for CYC1 is shown below as an example. The upper left pane of the table contains the position, DSM, e-value, and sequence at the putative processing site. The upper right contains a plot that shows an expanded plot of the DSM analysis of the sequence surrounding the putative plot. Five separate plots are shown, corresponding to the relative likelihoods of the four hexamer elements and the processing site: Blue: Element #1, Green: Element #2, Magenta: Element #3, Red: Processing Site, Gray: Element #4.

Position Score E-value Sequence
554 4.497125 0.09826345 ATTTCA AATTT

Probable element 1:
pos Seq DSM LL
502 TATTTA 4.75845 1.29864
511 TTTTTA 3.51116 -0.0282968
514 TTAATA 3.20118 -0.132134
506 TATTTT 2.96226 -0.639279
498 TCCCTA 2.69938 -0.430468
Probable element 2:
pos Seq DSM LL
530 TATTAA 4.16639 0.329585
528 AGTATT 4.14307 0.0886923
533 TAAGAA 4.09684 0.809699
522 TATGTT 3.89854 0.0133042
534 AAGAAC 3.80091 0.878333
Probable element 3:
pos Seq DSM LL
541 TTATTT 4.44134 0.941751
542 TATTTA 3.9135 0.354329
548 TATTTC 3.78742 0.991885
540 GTTATT 3.77261 0.540196
546 TATATT 3.63468 0.798813
Probable element 4:
pos Seq DSM LL
565 TTTTTT 3.99567 1.42378
563 TTTTTT 3.9721 1.42378
564 TTTTTT 3.93767 1.42378
566 TTTTTC 3.91987 1.29908
562 CTTTTT 3.84524 1.34893

The four lower panels list up to four (cut off at DSM=3) most likely sites for each of the four control elements in the DSM. For each listed putative element, the information given is sequence position, sequence, DSM, and the independent log-likelihood. The independent log-likelihood is the base 10 logarithm of the ratio of the probability of the hexamer being generated by the element model to the probability of the hexamer being generated by random yeast sequence. In contrast the DSM value gives the relative likelihood of the hexamer being used as the element within the context of the complete DSM. Comparison of the DSM and the independent log-likelihood allows investigation of the use of sub-optimal words as processing elements.