We have been urging the manufacturers of sequencing instruments to provide these values for several years and a number of independent groups have started to develop their own base calling algorithms which will also output numerical estimates of base accuracy. One way or another the numbers should be available in the near future.
As an interim measure to get the software developed we are calculating our own numerical estimates of base accuracy and storing them in the SCF files. The calculation performed is simply the area under the trace for the called base divided by the area for the next highest trace in the same position. The values are normalised to lie between 1 and 99 and the special values 0 and 100 are reserved for "hand edited" bases. Note at present we only store a value for the called base in the gap database but SCF files have space for 4 values for each position. Depending on the properties of the numerical estimates produced eventually by new base calling software we may need to switch to storing all four numbers in the gap database.
At present all we can say about the numbers that we currently calculate is that they are inversely correlated with the numbers of edits made. That is, the lower the value of our estimate the higher the number of edits that are made. These calculations were done by comparing the final edited versions of readings for completed projects with their original sequences, and then looking at the numerical estimates of base accuracy for the edited bases.
Note that until better accuracy estimates are available the consensus calculation is set, by default, to work in the old way. If users elect to try the use of our current accuracy estimates they will see that, within the contig editor, those bases that are above the accuracy cutoff appear in black letters and those below are red. The cutoff values can be adjusted using repeater buttons at the top left of the contig editor display.