Sample stability on 3700 deck. (and phred calling 3700 data)

Phillip San Miguel pmiguel at purdue.edu
Tue Oct 26 16:19:04 EST 1999



Paul Shinn wrote:

> Phillip San Miguel (pmiguel at purdue.edu) wrote:
> : Is there any consensus on how long samples can be left on
> : the 3700 (without covers--we haven't got the fang piercers
> : yet.) Could light be causing the sample degradation? Do
> : different sequence reaction chemistries cause different
> : effects here?
>
>     Currently, I use only 96 well plates so 4 plates is the most I can run
> at once.  I always run the big dye primers before the big dye terminators
> and I never run more than 2 BDPs in one sitting.  I have never been able
> to get good sequence from the 3rd plate of BDPs so I'm going to say from
> my experience that BDPs resuspended in 10ul 50:50 water:formamide will
> last 6 hours on the deck (time from hitting Run button to time of load of
> 2nd plate).  The last plate of BDTs is never as good as the first or
> second as evidenced by G dropouts.
>     I have been able to rerun a plate of terminators after it's sat on the
> deck over night but usually this only works if the reactions were strong
> to begin with.  Our normal shotgun chemistries are 1/8 BDTs and 1/4 BDPs
> off M13.  Both sets are resuspended in 10ul 50:50 water/formamide (EDTA).
> I'm using the modified Low Temp run module for terminators.  When we have
> good M13 templates, our read lengths are consistently well above spec--by
> eye at least.  I can't be 100% sure yet because phred isn't trained for
> 3700 data yet.
>     I'm excited about the fang release, too.  I'm even more excited about
> POP5, though!
>
>                                                         Paul
> [...]

Thanks Paul, this is just the sort of information that is useful.
By the way, results presented by Paul McEwan (Whitehead) at the 3700 users
meeting at the Genome Sequencing and Analysis Conference suggested that Phred
is just as good at estimating the quality of 3700 data as 377 data. It's
estimates were conservative for both. Specifically it gave lower-than-deserved
quality scores to later bases in a read.

Phillip San Miguel





More information about the Autoseq mailing list