first previous next last contents

Poor Quality

Read1      ACC*AGT*CGTA
Read2      ACC*AGT*CGTA
Read3      ACC*AGT*CGTA
Read4      ACCCAGTCCG
Read5      ACC*AGT*CGTA
Consensus  ACC*AGT*CGTA

This is identical to the first case, except we have two edits within Read4 in close proximity. This is usually due to a poor quality reading, which can be checked by examining the trace and confidence values. Whilst we could continue to make edits in the normal fashion it may be wiser to take another approach.

One technique is to adjust the cutoff data for Read4. By marking the data as hidden, this portion of the reading will no longer be used for producing the consensus. However we can only extend the cutoff data at one end or the other; it is not possible to have "hidden" data part way through a reading except by modifying its confidence. Note though that adjusting the cutoff data may mean that we have no data for one strand, which should be solved by extra experiments.

If the reading is poor quality along its entire length, then disassembly is also a viable option. Using Highlight Disagreements (editor Settings menu) or Check Assembly (see section Check Assembly) is a good way of finding such readings. Note that disassembling readings may have other implications. It could cause a hole in the contig (in which case it will be broken in two) or it could cause a single stranded segment. If this is the case the user needs to weigh up the work involved with making many edits along the length of this reading with performing another experiment to obtain better quality data.


first previous next last contents
This page is maintained by James Bonfield. Last generated on 2 Febuary 1999.
URL: http://www.mrc-lmb.cam.ac.uk/pubseq/manual/gap4_71.html