Here is a post i sent to ABRF, hopefully some others will
be intersted in this.
Seems to be a number of issues which are complicating the low
sensitivity issues most of us are experiencing using the dRHOD
chemistry. In summary:
-The concentration of dye in the terminator mix is lower.
-More template is needed especially using half reactions.
-The ethanol precipitation of dRHOD samples seems less efficent,
(efficency seems to be related to template ammount, if there is
lots of DNA it precipitates much more easily).
-There is considerable differences in sensitivity between instruments.
-Instruments with the XL upgrade may have lower sensitivity.
-The new mix is not as forgiving of poor samples as the old mix.
I spent 2 weeks of pure hell trying to sort out why 50% of our samples
were failing last month. This happened exactly at the same time we
switched to dRHODS, had a laser replaced, a burnt out backing plate
(2 months after our warranty expired of course!) and an XL upgrade
and it was a nightmare trying to figure why our signal strengths were so
low. It turned out that not only were signal levels low because of
the dRHODS and the XL upgrade but we had crap on our lens and only had
50% laser output after the lens. Such a simple proplem was so hard to
diagnose and the RAMEN sensitivity test ABI uses to test sensitivity
seems
to give highly variable results.
ABI really should have done more testing and given us more feedback about
what to expect both with the dRHODS and the XL upgrade especially after
charging us an arm and a leg for the upgrade to begin with!
While sensitivity is not as critical in sequencer performance as length
of read and high signal to noise ratios in the real world it is
frustrating to have 5-10 samples per gel with visible bands failing
to analyse. If samples have such a low signal there is usually a problem
with sub-optimal sequencing reactions but that is a side issue
if we have a situation where the sample may sequence fine on one
instrument and not on another! This is not a healthy situation.
I would really like to get a database of results from pgems run as
close to possible under identical conditions (i.e same gain, full
reactions, same loading ammounts, same number of lanes, same
length of read, same purification method etc) so we can have some
comparison between instruments to get a better handle on this.
I realise using average signal strengths quantatatively will be
inaccurate but at least it will be a guide and will help to identify
instruments with worse than average performance so we can then do
something about it. It would also be usefull to have a central point
to send chromatograms . Maybe we should keep the data anonymous so
the information cannot be used politically against sequencing rivals.
Specifically I would like to propose that we get as many instruments
as possible to seguence pgem at a gain of 4, full reaction, 2XS
7 hour run, loading 1ul/4ul on 48 lanes XL or 36 wells .
We can load the standard on 2 lanes, one a normal 1ul loading
and on the other load an equivalent of 200ul , diluting the
sample to 1:5 to reduce pippetting errors.
What do other users think? Any better suggestions?
I am going to try a 1/5 dilution out to see what we get this weekend.
My last dRHOD pgem (1/2 rxn, gain of 4, loaded 1 of 2ul, 48 well
8hour 2XS run) gave signals of G228, A236, C317, T192.
Finally some of my users have tried adding 1ul of a 5mg/ml
glycogen stock to there samples to help ethanol precipitation.
This has helped 2 users but it made no difference to 2 other users
,the only side effects seem to be more salt/blobs.
Anyone else tried glycogen?
___________________________________________________________
Macky Edmundson - Queensland Institute of Medical Research
Socrates - To Be is To Do
Jean Paul Sartre - To Do is To Be
Frank Sinatra - Do Be Do Be Do
email mackyE at qimr.edu.au (Australia)
_____________________________________________________________
