How good should automated sequencing be? - (Nov/01/2006 )
If I have a PCR product, and sequence it directly in an automated sequencer, both in the forward and reverse directions, and find that these two sequences are only 90% identical (instead of 100% identical), should I:
1. Try to optimise sequencing conditions to get 100% agreement between forward and reverse before submitting to genbank?
2. Take the consensus sequence between the two, stick in Ns everywhere that doesn't agree, and submit that?
3. Worry a lot about submitting something with potential error or not worry to much about it?
What is a reasonable degree of agreement in the forward and reverse directions on a PCR product? The product is 1 kilobase, and about 400 bases overlap from each end.
Normally, I'll have more trust on the initial 400 to 500 nt sequence. After that, they're more error prone, with base addition, deletion, more NN's etc....
I'd recommend primer walking to get to the 3' end of your 1 kb product. Or you could compare the overlapping region within 100 nt instead of 400 nt. You'll see that they'll be ~100% identical instead of ~90% when you included 400 nt.