Sorting standard operating procedure Summary

Elaine Kunze mek4 at psu.edu
Mon Mar 26 14:47:40 EDT 2007


Here is the summary of SOP for sorting in a core 
lab.  The opinions are diverse.  I loved reading 
all of them.  Many thanks for all of you who took 
the time to answer.  I really appreciate it.  Elaine

The questions: What do you routinely provide and 
what does the researcher provide:

1.  Do you count cells before and after sorts to estimate yields?
2.  How do you handle purities on rare event samples?
3.  Do you assess viabilities before and after sorting?
4.  Does your customer have input on flow rate, abort rates, etc?
5.  Do you routinely filter samples in the lab or is it done by the researcher?
6.  Billing policies on clogging samples, setup
7.  How do you handle timing?  Do you stay late or pull the plug?

**********
We do not have written policies apart from what we charge (actual time
on plus 1.5 hr set up/clean/decontamination with extra for CloneCyt),
perhaps we should, but in answer to your 
questions .This is for a FACSVantage SE

1. No that is up to the user, but I do give threshold, left, right and
abort counts from the instrument if asked for, but that can open up a
whole new can of worms (user and machine counts not matching up).
2. Luckily that does not come up for most if not all my sorts so no
input there.
3. No again that is up to the user,
4. No not directly but I will talk to my customers and try to give the
optimal service for what they want e.g. if they want every single cell
then I will sort slowly and they will have to pay more.
5. Filtering by the end user although I will sometimes re-filter if it
looks claggy when it get to me.
6a. Big "stopped dead needs a nozzle change" blockage, the sort is
cancelled and they pay up to the time of the blockage (plus
setup/decon).
6b. Semi block (stream goes off, droplets go funny etc) but can be
cleared without need to change nozzle, resterilize etc, they pay from
start to end of sort including the time taken to clear the blockage and
recalibrate/stabilise. Because it was their fault it blocked anyway.
Unless they decide to stop, in which case 6a applies.
7. I am in the ? fortunate? position of not being overly busy these days
so it does not happen often but when necessary I will stay late. The
latest I have finished sorting is 8 pm which means getting home between
10 and 11 pm (with an 8:30 am start) although I rarely start sorting
before lunchtime.
*******

)	Do you count cells before and after sorts 
to estimate yields?  We take e-mail requests in 
advance for booking time. We estimate 25-30 
million cells can be passed per hour. We ask for 
an estimate of the cell number to be provided to 
insure that they book enough time. If they book 4 
hours they shouldn’t have much more than 100 
million cells if they want them all sorted.  For 
estimating yield, we suggest that they determine 
the frequency of the population of interest 
first. Let’s say it is 25% of the total and they 
need 15 million cells in the end. Theoretically, 
they would need to bring 60 million cells. Since 
the sorter loses almost half in the process (more 
or less depending on the instrument) we ask that 
they double the initial number of cells.  In this 
case they would provide 120 million cells and need to book 4.5 hours or so.

2)	 How do you handle purities on rare event 
samples?  On the Vantage, we can’t really do a 
purity on less than 50,000 events.  The Aria 
always gets a purity, even on a few as 5,000 cells.  And it’s usually

3)	 Do you assess viabilities before and 
after sorting?	We ask that they do it.  It’s 
their sample!  If we note a number of dead/dying 
cells by FSC/SSC then we tell them that they are already there.

4)	Does your customer have input on flow 
rate, abort rates, etc?  We listen to their 
requests then advise as to whether it is too slow 
or too fast to either finish the sample or to get 
viable cells in the end.  Some of the cells 
(tumor, DCs, myeloid) hate to be run too fast. T cells can always take it.

5)	Do you routinely filter samples in the 
lab or is it done by the researcher? We ask that 
the researcher provide us with a sample filtered 
through a 70um filter. Still, we check before the 
sort and if the sample wasn’t filtered or is 
clumping we will filter it ourselves at the instrument.

6)	Billing policies on clogging samples, 
setup. Most people’s templates are already in the 
instrument but if not then they get set up at the 
start of the sort time and the researcher is 
billed for that time. No one else can use the 
instrument and the lasers are rolling so I say 
bill it!  Same for clogs. Sheath doesn’t plug the 
machine.  If they provide a sample that plugs it 
up then they get to pay for your time to unplug 
it.  That always reinforces the merits of filtering in advance!!

7)	How do you handle timing?  Do you stay 
late or pull the plug? On all of our instruments 
we have a policy that you can run as long as the 
next user will let you. If someone is booked 
until 2:00 and they need another 30 minutes, it’s 
up to the person who is on at 2:00. But that 
person can’t expect the next to be as 
generous.  For sorting, we have a dedicated 
operator on the Vantage and it runs from 
9:00-5:00 every day. Longer sorts must be 
scheduled in advance, otherwise they cut at 5:00 
p.m. If the researcher is just late and/or 
unorganized we won’t stay longer. The truth or 
the matter is that you have to define your own 
boundaries, otherwise they will have you there 
until 10:00 in the evening because they didn’t 
want to get up before 11:00 a.m.!  Trust me, we’ve seen it!
******
1.  Do you count cells before and after sorts to 
estimate yields?   NO -- BUT WE TELL OUR USERS 
THAT THEY SHOULD DO THIS.  THEY OFTEN DO NOT DO 
IT,  AND THEN WE TELL THEM THAT WE CANNOT TALK 
ABOUT CELL YIELD IF WE HAVE NO IDEA HOW MANY 
CELLS THEY HAVE BROUGHT TO THE INSTRUMENT.  THE 
OTHER THING TO BE SCEPTICAL ABOUT IS THAT MANY 
USERS COUNT THEIR CELLS *BEFORE* THEY STAIN 
THEM.  SINCE MANY CELLS ARE LOST DURING STAINING 
AND WASHING,  THIS COUNT DOESN'T DO ANYONE MUCH GOOD.
2.  How do you handle purities on rare event 
samples?  WE JUST LOOK AT THE SORTED CELLS AS A PERCENT OF THE TOTAL.
3.  Do you assess viabilities before and after 
sorting?  WE DO NOT DO THIS.  SOME OF OUR USERS 
DO, AND WE RECORD THE INFORMATION.
4.  Does your customer have input on flow rate, 
abort rates, etc?  WE TRY TO GET THE USER TO 
COMMENT ON FLOW RATE WITH REGARD TO HOW MANY 
CELLS THEY ARE WILLING TO LOSE TO ABORTS.  WE 
HAVE A SPREAD SHEET THAT HELPS THEM TO MAKE THIS DECISION.
5.  Do you routinely filter samples in the lab or 
is it done by the researcher?  IT IS DONE BY THE 
RESEARCHER,  BUT WE HAVE FILTERS IN THE FLOW LAB 
TO USE IF THE SAMPLE SEEMS TO BE BLOCKING THE LINES DURING THE SORT.
6.  Billing policies on clogging samples, 
setup  WE BILL FOR ALL USED TIME (BUT DO NOT BILL 
FOR TIME WHERE IT IS OBVIOUS THAT INSTRUMENT 
MALFUNCTION HAS OCCURRED).  WE ROUTINELY BILL AN EXTRA 1-2 HOURS FOR SETUP.
7.  How do you handle timing?  Do you stay late 
or pull the plug?  OUR SORT OPERATOR IS VERY 
GENEROUS AND HE USUALLY STAYS LATE, IF 
REQUIRED.  HOWEVER,   WE AIM TO MAKE A GOOD GUESS 
AT HOW LONG THE SORT WILL LAST IN ADVANCE.  IF 
THE OPERATOR REALLY NEEDS TO LEAVE AT A CERTAIN 
TIME,  THEN HE WILL TELL THE USER IN ADVANCE THAT 
HE DEFINITELY NEEDS TO LEAVE AT 5PM OR 6PM OR 
WHATEVER.  WITH OUR ABILITY TO SORT AT HIGH 
SPEEDS,  WE CAN OFTEN SPEED UP SORTS BY 
CENTRIFUGING THE CELLS TO CONCENTRATE THEM.
***
We are a smaller core facility and while we don’t 
sort every day we are sorting most days of the 
week for multiple users.  Here is my two cents:

1)	 I don’t routinely count cells before or 
after, I leave that up to the investigator.  That 
said, if I felt that I needed to do that for my 
own benefit, I do have the capability of 
implementing it.  With some investigators it 
might be a good idea.  While I don’t always think 
the sorter (FACS Vantage DiVa) is completely 
accurate in its sort count, feel pretty confident about the threshold counts.

2)	  For really rare populations I don’t 
check for purity afterward.  What we have done as 
a kind of work around for that is to first sort 
another population of similarly stained cells and 
check them for purity.	It’s not perfect, but at 
least we can tell how well the machine is behaving.

3)	 Occasionally I will assess viability 
before, mostly with new clients or people who 
routinely have unhappy cells.  Usually I don’t 
assess viability afterward but leave that for the 
investigator.  Again, if this is a problem, it 
might be better to be more rigorous in this area.

4)	 In general, no.  I usually set the flow 
rate to minimize the abort rate and maximize the 
efficiency as much as possible.

5)	  Filtering is mandated, but is done by the investigator.

6)	 At the moment we do not charge for setup 
time. If a sample clogs in the middle of the sort 
and I need to restart, etc, I usually do not 
subtract that time from their charges, so they do get a bill for it.

7)	 I have stayed late to finish.	It’s a 
judgment call most times.  Lot’s of things go 
into that decision.  I don’t make a habit of it, 
but depending on circumstances it might happen.
****
I am an operator using a BD FACSVantage with DiVa 
and am the only person running the machine. I do 
not do research or provide staining services, I 
simply run the sorter all day. I do training and 
QC on the analytical machines and light repairs 
when needed as well as consult for staining and 
troubleshoot. I do not provide antibodies, 
controls or even sample tubes like other core 
facilities. I do provide FCS analysis software 
that is located on the university network.

1.  Do you count cells before and after sorts to estimate yields?
Before a sort: No. But users who count will 
always count before they do anything to the 
cells, like wash 3 times. In this case they will 
usually bring anywhere between 10 and 80% of what 
they started with, and therefore grossly over 
estimate the number of cells you should be giving them back.
After: The machine usually tells you how many 
cells you sorted and you should check your 
machine to make sure this is accurate. Mine is 
spot on so I don't count afterwards.

2.  How do you handle purities on rare event samples?
I check purities on non-rare event samples by 
sorting that last 100ul (or ~10,000 sorted 
events) into a separate tube, rinse with water, 
then put that tube back on the sorter, take a 
data file of the purity and give the user the 
print out of the purity. For rare events I don't 
because it will take forever and waste their 
cells. If you expect a user that day to have a 
rare population, you can sort a mixture of beads 
(I use calibrite beads from BD) and do a purity 
check this way. Users generally assume that I 
have the machine set up properly and many don't 
care if I do a purity check or not, but others 
expect it. When I worked for Fred Hutch in 
Seattle, we never did a reanalysis on cells, we 
just did a bead check twice a week. I prefer to test real cells when I can.

3.  Do you assess viabilities before and after sorting?
I always request users to add PI to their 
samples. Gating conservatively on FSC/SSC will 
help a lot if they don't. If viability is in 
question post sort, I will add PI to the purity 
tube after I have collected 10,000 cells or so, 
because it gets washed out after the first sort. 
As a side note, if your sorter has chilled 
collection and sample chambers, this helps viability immensely!

4.  Does your customer have input on flow rate, abort rates, etc?
No (although some think they should). The cells 
and machine set up really determine this. This is 
the operator's job, to become familiar with the 
combination of the set up (ie, nozzle size, 
pressure) and cell type/experiments and the resulting output/efficiency/speed.

5.  Do you routinely filter samples in the lab or 
is it done by the researcher?
I ask that users filter their samples before they 
bring them to me, but I do it also if it is 
needed. Many times if the cells sit too long they 
will need to be filtered again. I use the 5mL 
Falcon tubes with the blue cap filter.

6.  Billing policies on clogging samples, setup
I don't bill if the sorter is clogged and their 
sample is contaminated and I cannot recover the 
cells. However, most of the time I am able to 
resort the contaminated tube very quckly and efficiently, in this case I bill.

7.  How do you handle timing?  Do you stay late or pull the plug?
This has been a topic of recent discussion. Since 
I am the only one running the sorter, I feel 
obligated to stay and finish the job. Many days I 
miss lunch. I keep a flexible schedule and don't 
keep to strict regular hours although I try and 
end sorts at 5 because 6 is when I will pull the 
plug and this gives me a little room to stay if I 
need to. If I can I might come in late the next 
day. Other labs do stick to a strict schedule and 
this is probably better to be honest. People tend 
to take advantage of an operator's willingness to stay late.
******
This info is for a core facility with > 8 
sorts/week on a Beckman-Coulter Altra.

Core can provide the customer with sterile tubes 
(Falcon 2054, 2063), Flacon filter tubes (2235) 
or filter mesh if they do not have or can't 
borrow from another investigator. If a new 
application is requested the core will purchase 
the initial reagents to expand our services. 
Specific filter sets are purchased by the core, 
i.e. GFP/YFP, Hoechst Blue/Red (side-population), etc.

1. Customer performs all cell counting and is 
strongly encouraged to do so rigorously. If a 
vast discrepancy is observed, i.e. they say they 
are providing one million cells but is off by a 
factor of 10X due to miscalculation operator will 
count. Post sort number of drops (and aborts) are 
provided and >95% of the time they match the post sort cell numbers.

2. Post sort analysis for purity of rare events 
are not provided (they hate sacrificing yield to 
verify purity) but a negative or non-rare 
population is always sorted to verify instrument 
set up and drop delay is correct and demonstrate 
potential purity, typically 98%. All post sorted 
cell surface immunophenotyping samples are 
required to be re-labeled, in the case where 
cells are live (unfixed) and may lose some antibody during the sort.

3. Customer is responsible for viability 
assessment, pre and post. (Over-processing 
tumors, sort time and temp can be a sticky-wicket 
to argue with a customer). Always recommended to 
use a viability exclusion dye in the sort such as 7-AAD, PI, etc.

4. Flow rates are determined by the application 
and sort frequency (1/4 max of frequency, 80KHz 
has a max rate of 20,000 events/sec). If a large 
cell is sorted using a 140um tip which creates 
larger volume drops the sheath pressure is 
between 6-8psi and the frequency is between 
8-12KHz for a rate no greater than 3,000 
events/sec. Abort rates help to determine if the 
tip is damaged or if the cell prep is sub-standard.

5. Customer filters samples.

6. The customer gets 1 free tip (~ $1,000/ea). 
Upon replacement they pay the replacement cost 
for the next one lost. Our service used about 
twelve tips last year. In our service we provide 
a separate tip and tubing for each customer, or 
sometimes for each two customers. One customer 
has a 100um and a 140um tip. Tips and tubing 
changed out 2-3X daily. Billing initially was a 
minimum of 2 hours at $70.00/Hr but since we are 
sorting so often the 2 hour minimum charge has been waved.

7. Stay Late.
******
1.  Do you count cells before and after sorts to estimate yields?
NO.  If I get a serious hassle I don't trust 
manual cell counting, but I will use an automated cell counter.
2.  How do you handle purities on rare event samples?
I take advice from the researcher as to whether 
they want to lose 2-5,000 events, mostly I trust 
my instrument as my researchers want every single 
cell they can get!  If your instrument is good 
and your QC thorough you don't get too many 
hassles other than major board failures.
3.  Do you assess viabilities before and after sorting?
NO.  Researchers are free to add a viability dye if this is important.
4.  Does your customer have input on flow rate, abort rates, etc?
To a very small degree.  I run the sorter at its 
optimal settings, maximising yield, purity and 
recovery.  If a researcher reports low cell 
numbers and lots of debris I will run at a lower 
pressure to see if that helps.	I will advice on 
my current instrument that yields are increased 
with tow different two way sorts instead of one 4 
way in some rare event cases.  I generally find 
that if the researcher directs the instrument 
operation the results are very poor.  I am 
employed as the cytometry expert, I work the 
instrument every day, I am supposed to know what I am doing.
5.  Do you routinely filter samples in the lab or is it done by the researcher?
This is the researchers responsibility, although 
in the past I have supplied the filters to make sure the job is done.
6.  Billing policies on clogging samples, setup
Billing is from when the sort starts (or first 
control is collected) (or in new cases when I 
start to build the sort layout).  Clogs are 
usually because the researcher has failed to 
filter - their problem not mine.  At the moment I 
don't charge for set up.  In other labs I charged 
a minimum for in house and 1 hour for external clients.
7.  How do you handle timing?  Do you stay late or pull the plug?
I stay until the job is done.

***
1.  Do you count cells before and after sorts to 
estimate yields? Yes, also,  we take counts 
during sorting electronically to get a sense of 
how many cells have gone through.
2.  How do you handle purities on rare event 
samples? I am not sure of what you are asking 
here, but if we have a rare event sample that 
needs sorted, we use an enriched sort to enrich 
the population (80%).  Some people we sort for 
request high purity, which is lower yeilding.
3.  Do you assess viabilities before and after 
sorting? Yes, we do this using microscopy, some 
samples after sorted are beaten up a 
little.  Before hand we use microscopy and trypan blue method.
4.  Does your customer have input on flow rate, 
abort rates, etc? Yes, but we find that typically 
they let us decide what is best.  In fact, I 
haven't had one person make the decision.
5.  Do you routinely filter samples in the lab or 
is it done by the researcher? We usually sorted 
cells freshly isolated from tissue.  Yes we 
filter them using Falcon tubes with filters on the caps.
6.  Billing policies on clogging samples, setup 
Our policy although, I don't agree with it, is 
that we don't bill for downtime due to 
clogging.  So if it clogs the time stops until we 
are ready to sort again.  I think this should be 
something that the user should understand is a 
possiblity, especially working with fluid 
systems.  Also, if samples are being sorted into 
something like Trizol for subsequent RNA 
isolation, and a clog were to contaminate the 
sorted material, we don't charge for the 
experiment.  We don't charge for general setup, 
but we do charge $45 for sterile setup.
7.  How do you handle timing?  Do you stay late 
or pull the plug? We stay late and rarely pull 
the plug.  Pulling the plug for us indicates 
there are no positives to be found.  It can mean late nights.
****
1.  Do you count cells before and after sorts to estimate yields?
The before sort counts done by the researcher. 
Mainly for them tell us the concentration of the sample.

For my FACSAria I cannot control the event rate 
so that is determine by the concentration of the sample.
Low concentrations will not sort optimally and 
high concentration tend to clog the machine.

The after sort count is determine by the sort 
counter on the machine during the sorting process.

2.  How do you handle purities on rare event samples?

We set sorting masks for different sorting templates.
BD's Purity mask is working well for us when we 
do SP positive cells which are as low as 0.1% of total cell events.

3.  Do you assess viabilities before and after sorting?

Depends on the researcher. Not all of them are 
concern with viability, if they are getting the 
cells for RNA work then they won't need to know 
it. Viability is assess through vital dye 
staining (PI, DAPI, Hoechst). Before sorting 
viabilities are determine through analyses of the 
samples, and after-sorting viabilities are 
determine through analysing the post-sort tube.

  4.  Does your customer have input on flow rate, abort rates, etc?

If they have previous sorting experience with 
their cells and they know what works best for 
them, I'll let them have their preferences. If 
they haven't, we provide them with guidelines.

5.  Do you routinely filter samples in the lab or 
is it done by the researcher?

This is done by the researcher. It is their 
responsibility to provide us with single cell suspensions.

6.  Billing policies on clogging samples, setup

We charge by the hour. Inclusive of time taken to 
set up templates or to recover from clog 
downtime. If machine error occurs during sort 
that is not customer's fault, we will not charge 
for time taken for troubleshooting. Some 
facilities charge additional first time setup fee, but that is your decision.

7.  How do you handle timing?  Do you stay late or pull the plug?

Customers call us prior to coming down to book a 
time slot with us. We sometimes have to stay late 
due to unforseen circumstances. But we try to set 
a limit to the timeslot they can book, like they 
can only book the machine until 5pm, we give 
ourselves the remaining 1/2 hr of the day to 
clean up and then pack up at 5:30pm.
*******
1)	 NO, that is the investigator’s 
responsibility. If they have a complaint you can 
certainly verify their counts the next time.

2)	 That’s a tough one; it is really the 
investigator’s call as to whether they want to 
sacrifice some of the cells or not. I suggest 
they do it for one or two experiments and as long 
as the assays they perform with the sorted cells 
have similar results between sorts then I think 
it is safe to conclude they share similar quality in purity.

3)	 I always recommend the use of viability 
dyes in samples to be sorted. This will at least 
give you presort viability. If they choose they 
can restain the post sort sample for viability as 
well. However, again, this is the investigator’s responsibility.

4)	 I explain the nuances of aborts, sort 
rate efficiencies, and impact on yield and 
recommend certain flow rates to achieve a minimal 
of 80% efficiency. However, if an investigator 
insists we go faster understanding the impact on 
their yield, that is their call.

5)	 We recommend all samples be filtered 
prior to sorting. Occasionally there may be 
subsequent clumping and we have filter material 
available to refilter if necessary. However, I 
make them filter their own samples. If they screw 
it up it is their consequence, however if you 
lose their sample in the process of filtering that is a bit more disastrous.

6)	 We routinely charge an hour set up time 
prior to sorting. If there are multiple sorts in 
a day without an instrument configuration change 
between, thy set up time is shared equally 
between investigators. If a sample clogs the 
instrument the investigator is still on the clock 
until resolved. This does help insure they filter their samples in the future.

7)	 We do not punch a clock, however, if an 
investigator schedules 2 hours late in the 
afternoon and brings enough samples to sort for 4 
hours then I inform them up front that we will 
only get to half the samples. It is a good idea 
to consult with investigators prior to their 
scheduling of sort time to make sure they have a 
realistic idea of what can be accomplished in “X” 
amount of time. You should have a fairly good 
estimate of throughput of cells/hour using 
various nozzles and pressures. It is also a good 
opportunity to make sure they understand how to 
prep their cells for sorting as well as to make 
sure they know what controls to bring. When we 
are dealing with back to back sorts, this can 
sometimes be problematic as things do not allows 
go according to plan. For this reason I like to 
add 30 minutes or so of cushion time to the sort 
schedule between sorts, but even then we can 
sometimes run into unforeseen problems. Everyone 
has to learn to be a little flexible in these instances.
	 Overall we subscribe to the philosophy 
that we are here to provide expertise and 
education as to the ideal way of doing things. 
Ultimately, however, it is the investigator’s 
responsibility to insure optimal experimental 
conditions. There will be those that insist 
things be done a certain way and in doing so do 
not get the best bang for their buck. However, it 
is your (or our) responsibility to make sure we 
educate them on the potential consequences of 
their decisions. I draw the line when their 
decision generates erroneous data. The comment 
that irritates me most is the “but I have always 
done it this way”. Just remember education is 
your best tool to insuring cooperation. Some 
people have just learned to do things incorrectly 
and need to be educated on how to do them properly and why.

We have a tough job but it often has its rewards!
******
1. We don't do cell counts. It's up to the PI's lab to prepare the sample,
and that includes doing a cell count to accurately set the concentration
of the final suspension. After sorting, they can either take the sort
counter as an estimate or count themselves if they need something more
accurate. If we had a Coulter counter it would be 
no big deal to do the counting
ourselves. Of course that's one more mouth to feed and the costs
would have to be passed on to the clients.

However, it is helpful to calculate yields internally to establish
optimal sorting conditions and quality control the sorter. We did this
jointly with the director's lab many years (and sorters) ago in which
they did the counting. (We found out that for their experiments with our
EPICS 753 the count in the collection tube was always about 80% of the
sort counter.)

Also, you gain some peace of mind by doing your own counts. Although it
is easy to qualitatively state that today's sample concentration is lower
than yesterday's by virtue of the cell rate on the sorter, a cell count
is more definitive. A low yield in this case is due to sample prep, not
the sorter. If under constant conditions the sorter one day produces
a collection count much lower than the sort counter, then there was a
problem with the sorter (or the set up).

If it is not too much bother I would recommend you do cell counts at
least for internal purposes. If you decide to disclose them to your clients
then they will become mandatory for all samples. I would recommend using
an automated cell counter.

2. For rare event samples my approach is to measure the purity on a quick
bead sort before and after the rare event sort under the exact same
conditions (except initial fraction of target cells). I realize this is
"cheating" because I'm using beads, but without SweetSpot technology
the breakoff can drift over time. I'd consider this to be the biggest
source of impurity if the target population were well resolved. Of course
if there is a clog, all bets are off. I also have more confidence if
machine settings didn't need to be tweaked during the sort.

For rare event detection (not sorting), we sometimes do these on the sorter
and sort onto slides in an attempt to verify the detection. If you can
spare enough cells, you could sort some cells onto a slide. The precision
of the purity estimate would depend on the number of cells. Ideally you'd
want to sample at the beginning, middle and end of the sort, but even with
only 20 cells per sample that's asking for a lot of cells when it comes to
rare event. Assuming as above there were no clogs or tweaking, I'd have
some confidence in sampling only at the end of the sort. (The beginning
should be fine because you took special care to assure the sorter was
working properly.) Having access to an image or scanning cytometer may
make this easier.

3. I only qualitatively assess viabilities prior to a sort based on
scatter. If it looks bad, i.e. < 85% viable, I'll inform the PI that
the sort will likely go poorly; it's up to them to decide if they want
to proceed. Most do proceed, and the sorts do mostly end up poorly.

Again, it is up to the PI's lab to measure viability during sample
preparation. If they need to measure it again post-sort, that's up
to them. Not everyone is concerned with post-sort viability.

Most PI's have been satisfied over the years with sorting without a
viability probe like PI. Afterall, PI is slightly cytotoxic, so why add
one more stress to the cells? In those cases where I find it helpful to
add a viability probe I usually use calcofluor white because it excites
in the UV and does not interfere with 488 nm-excited fluors which make
up the majority of our sorts. (Obviously for SP the protocol calls for PI.)

4. I haven't had a customer with more sorting experience than myself
(going on 17 years). They usually defer to my judgement. Some like to
set their own sort gates. I'm fine with this since that makes them
responsible for purities. Typically flow rate is determined by the
parameters the PI is really interested in: acceptable purity, number of
collected cells needed, time, and number of cells provided.

After a terrible experience trying to appease a PI by training her techs
to run the sorter I will never allow this again. Nobody is allowed to
operate my sorter other than me except for current facility staff that
have been trained! I was in a pickle over this when my assistant left the
facility to join another lab, and an allowance was made. (He's since moved
on and it is no longer an issue.) I don't care if your name is Herzenberg;
you're not touching my sorter!

5. We ask the PIs to filter their own samples. If a sample clogs the
sorter it (and any remaining samples) are automatically filtered. I
don't like to do this because it is a potential source of contamination.
I always inform the PIs of this risk, as in, "I cannot guarantee a
contamination-free sort if your sample clogs my machine, so be sure
to provide extremely clean samples."

6. Due to traditional practices, we do not charge for setup. Although
we've returned to mostly all-day sorts, there was a time when short sorts
were popular. At two hours to set up and half an hour to shut down, we
were on the short end of the stick when the only sort that day went for
5-15 minutes. As I said, our sorting has declined over the years, and
the biggest complaint is always that our services cost too much. After
25+ years of operating this way it would not bode well to start charging
for setups.

For clogs, the clock keeps ticking. Since the responsibility is put on
the PI to filter and otherwise prepare clean samples, they get to pay
when the machine clogs.

7. Again the tradition we've established is to stay late. It's up to
the PI to pull the plug. Most are reasonable about it. They've put a
lot of money and effort into their experiment; if the difference between
a success and a failure is 30 more minutes it seems a shame not to give
them the time. (Not to mention it can be bad for business.)

My family would rather that I always leave promptly at 5pm, but the
nature of flow cytometry is such that 10-6 makes more sense than
9-5. We don't have the kind of user base where I can dictate the hours
and make all the techs work 7-3 or 7-7 instead. They want to work 9-5 too.
For sorts they usually come in early, so it's up to them if they want to
stay late(r).


II. IDEAL ANSWERS

In a perfect world, my answers would be:
1. Yes, with an automated counter.
2. No change with above.
3. Yes, with my own choice of viability marker as above.
4. No.
5. Yes, with the expensive filter-in-the-cap test tubes, billed to the PI.
6. All billable time is billed, including setup and clogs. (I'd cut them
a break and charge our lower rate that just covers personnel cost instead
of the higher sorter rate for setup and shutdown. For multiple sorts in
one day [which has become rare for us], these costs would be split.)
7. The facility hours are set in stone, including a half hour shut down
time on all cytometers. All staff leave at 5:00pm, so all cytometers get
shut down promptly at 4:30pm. No exceptions. We'll still stagger lunches to
handle long sorts, but otherwise the facility is closed between noon-1pm for
lunch as well.

Unfortunately, even in a perfect world, the politics of academia are such
that someone is bound to bend your rules.


III. OTHER CONSIDERATIONS

Some other things to consider are keeping extremely detailed notes on
every sort. Every single machine parameter needs to be recorded, including
exactly which 90 micron nozzle was used. I'd even go so far as to track room
temperature and humidity, and make this an issue with building services
such that they maintain acceptable levels of both.

Speaking of building services, our sorter used to be located in a building
with sick building syndrome. We literally had large particles -- not just
dust -- blown in from overhead ducts. We tried to use an air filter over
our duct, but it was no surprise when we still saw fungal contaminations
from time to time. (The black dust/particles were fungal spores from
organisms living in the ductwork and/or being drawn in from the building
airtake.) After moving to a new building the contaminations were always
bacterial -- no more fungus -- and even that was determined to be coming
from the "unique" fluidics of the FACSAria we had coincidentally started
using (which thankfully I no longer have to manage).

So you need to control for contamination. I'm sure others will provide
you with more details than I can. Going back and forth with PIs over who
is the source of contamination is a headache best avoided. It's even
worthwhile to have your own incubator just for this purpose.

Be sure to document all of your QC work. At the very least each sorter
needs to be QC'd every day it is used. Record all AccuDrop parameters,
and keep FCS files of all control particles (alignment, AccuDrop, Rainbow
. whatever else you use). Be overly thorough. This gives you a firm
footing when a PI accuses the machine (or you) of not giving the desired
results from run to run, when in reality the fault may lie in variations
in sample preparation.

*****
>1.  Do you count cells before and after sorts to estimate yields?   Always.

>2.  How do you handle purities on rare event samples?

In most cases, we do an enrichment step, either on the sorter or by
other separation protocols (bead columns, MACS, etc.).	We have found
that our purities are well above average in our post-sort tests (15 -
60% pure after enrichment, 99% pure post-sort).

>3.  Do you assess viabilities before and after sorting?

Although not always a prerequisite, we do ask our clients if they
check for viability pre-sort.  We have taken to using Propidium
Iodide during and after some cell-line sorts to establish actual
purity rather than "alleged" purity.  What I mean by this is that we
have seen some cell line post-sort samples look awful in post-sort
tests, some as low as 85% pure.  However, when we stain with a
viability dye post-sort, the contaminating events light up as PI+.

>4.  Does your customer have input on flow rate, abort rates, etc?

Yes and No.  "Yes" in that we do our best to educate our clients in
regards to how the sorter works, and "No" in regards to allowing our
clients  to dictate to us what is acceptable and what is not.  A very
high abort rate usually can be handled by an experienced sorter
operator so that the best purity and recovery can be obtained.	The
client does not necessarily have to have input in this regard,
although they are informed.

>5.  Do you routinely filter samples in the lab or is it done by the
>researcher?

Not "routinely"!  ALWAYS!!!  We always filter samples, almost always
just before the samples are sorted.  Usually, our clients do the
filtering at the sorter bench.

>6.  Billing policies on clogging samples, setup

We bill for 30min setup, no exceptions.  We don't bill for clogging.

>7.  How do you handle timing?	Do you stay late or pull the plug?

Our philosophy has been and will always be:  We sort until we are
done.  Our pre-sort discussions with our clients always give us an
excellent idea as to how long the sort will take.  This is, of
course, assuming that the sorter operator is an experienced hand and
knows what questions to ask!  We leave it up to the client to decide
when to pull the plug.	We are again assuming that the sorter
operator knows his/her business and can plan their days accordingly.

Over the years we've been running this facility, we've had impressed
upon us the impact our facility has on our client labs.  The
thousands of dollars spent on mice, monkeys, rats, humans, cell
lines, bacteria, viruses, reagents, other materials, not to mention
the time involved in preparing for a sort...it can get pretty hairy
from that perspective.	Add to that the precious time devoted by our
clients, be they graduate students, post-docs, lab technicians,
etc....again, our facility's impact has enormous consequences.	A bad
sort can set-back a grad student one year.  To that end, we do not
take lightly our responsibilities to our clients.  Our policies have
evolved and been designed around these and other principles.
****
    * They count their own
    * They have to take it on faith unless they 
are willing to give up at least 1000 cells
    * If at all possible, we add a viability stain to the sort
    * No – we can discuss whether the sort is for 
purity (as it is 99% of the time) or enrichment 
but we are the sort ‘experts’ - I don’t really 
care what they have read or how they did it where 
they used to work because they generally don’t 
really know what they are talking about
    * I do this – I have sterile 50 micron filters and a bio-hood
    * I am so lucky here – I do not bill
    * We discuss how long the sort will take 
based on: 1) how many cells they will need to 
have returned from an unlimited source or 2) how 
many cells they will bring from a limited source 
(in which case the whole sample will be 
sorted.)  If, for instance, we decide on 2 hours 
then we agree on a time slot – say 10-noon.  If 
they show up at 11, then they only get a 1 hour 
sort. If they show up with twice the number of 
cells we discussed, I will do as many as I can in 
the allotted time and give them back the ones I 
did not have time for.	If they told me the 
population they want is 5% and it turns out to be 
0.5%, then I will do what I can in 2 hours but 
they will not get as many cells as we had 
discussed. If I have trouble with the sorter and 
I cannot start until 11, then I will sort for 2 
hours; everybody else the rest of the day will be 
late with my apologies. So it is basically ‘pull 
the plug’ unless the delay is my fault. I think a 
core lab sorter has to be handled very 
differently from a lab-dedicated sorter so you 
can maximize the sorting efficiency and the time 
the sorter is in use. You cannot inconvenience 2 
or 3 or 4 customers a day because one user is 
late. You also cannot expect your staff to be 
kept late and then be back early the next morning 
for another full day of scheduled sorts.
    * ****
1.  Do you count cells before and after sorts to estimate yields?
No, the researcher has to give us an estimate 
when he applies for sort time. Based on the 
estimate of cells he is gonna bring and the 
number of samples the researcher is given a time 
period for sorting. If he brings more cells than 
he said he was the sort can be cut off at the end 
of that time period. Also in form of an example, 
if a researcher tells us 2 samples with 100 
million cells total he can get 3 hours, if the 
researcher consistently brings far less in stead 
of his estimate, say 20 million and uses only a 
fraction of the time allocated to him, that will 
be taken in account too the next time he applies 
for sort time but that is getting off topic.

2.  How do you handle purities on rare event samples?

We ask the researcher if he wants absolute purity 
and trade in some of those rare events or wants 
all of the rare events but with the chance of 
impurities. It really is up to the researcher

3.  Do you assess viabilities before and after sorting?

Up to the researcher although they very rarely 
ask for post sort viability checks, they often 
enough put in a viability dye before sorting.

4.  Does your customer have input on flow rate, abort rates, etc?

Yes they do to a certain extent but there are 
very few cases that a researcher wants to 
influence the speed, most of those few cases want 
to go slow, nearly always because they think 
their viability will go up or just that the cells 
find it more pleasant going at a slower speed. 
Usually when we explain that it doesn't really 
matter much if the cells are pushed through on a 
MoFlo with 60.1 psi or 60.3 psi, it is still 60 psi.

5.  Do you routinely filter samples in the lab or is it done by the researcher?

We recommend filtering at all times. If they do 
not filter and it clogs up the machine we filter 
it for them, this is stated in our sort policy 
that that is a perogative of the operator to do so.

6.  Billing policies on clogging samples, setup

Setup and clogging will be included on billing 
time. Setup does not take very long often enough 
we have templates for the researcher on file and 
the machine is already set up in the morning and 
we do not bill the researcher for that part.

7.  How do you handle timing?  Do you stay late or pull the plug?

Often enough we pull the plug on samples if it is 
more cells/samples than the researcher told us in 
his booking, but if it is a precious patient 
sample we might be pursuaded to stay over time or 
find another way to sort it for them. Pulling the 
plug on people is needed here since this flow 
core is a very high throughput facility. A single 
operator can have 3, 4, 5 or even more sorts a 
day of course depending on what the researchers booked.

*****
1.  Do you count cells before and after sorts to estimate yields?
I use an experiment sheet where people have to 
answer important questions about their sample. 
Cell concentration is one of those questions. 
Whatever method of cell counting they use will 
produce a different count than the cytometer’s. 
Flow experts have different ways to deal with 
this, but tell you users that their, if they have 
it, would be half or twice as many events as the 
instrument should count. It is easy to stick to the counts of the instrument.

3.  Do you assess viabilities before and after sorting?
If you can use a viability dye in the sample 
which is to be sorted, you would be doing yourself a great favor.

4.  Does your customer have input on flow rate, abort rates, etc?
I find that people who operate the machine have 
the best understading of how all of this 
parameters would affect the outcome of the sort. 
I talk to my users as I set up the machine so 
they can have realistic expectations for yield and purity.

5.  Do you routinely filter samples in the lab or is it done by the researcher?
Both, somebody will forget to filter at some 
point. Also, some samples might clump again after filtration.

6.  Billing policies on clogging samples, setup
You might want to schedule a setup experiment 
when working with new samples. Something small. 
Talk to the customer about the quality of the 
sample during setup, provide input into how to 
make things better. It took one of my customers 6 
months to optimize their cell dissociation 
protocol. Now, the samples never clog.
Hope this helps.
****
We do not estimate yields but our instrument 
gives a readout as to the number of cells sorted, 
so this info is passed on to the 
researcher.  They also receive the printout of 
the acquired data giving the % positive etc.

On rare event samples we usually do not perform 
purity checks as the researchers are more 
concerned with recovery of cells than the purity.

We do not assess viability before or after 
sorting, generally the samples are prepared and 
brought to use promptly and once we have 
completed the sort (into the media they provide), 
the researchers pick up the samples 
promptly.  Feedback from different researchers 
shows that recovery of single cell and bulk sorts are successful.

Regarding input on flow rate and abort rate, 
there is discussion as to what the priorities are 
for the researcher i.e., recovery of cell numbers or speed of sort.

We routinely filter samples and visually inspect 
them prior to running or sorting to ensure there 
is no clumping.  We also have the researchers 
give us all samples in 2mM EDTA/PBS to reduce the 
clumping possibility.  Despite these precautions, 
we still have clogs, some of which are major clogs causing hours of downtime.

For the billing, we do not bill for setup or for 
major clogs (usually the sort is aborted at that 
point).  If minor clogs occur during the sort, 
the entire sort time is billed for.

We schedule appointments for sorting.  There is 
discussion as to what the priorities are for the 
researcher and the amount of time available.  If 
live cells are required (for subsequent culture) 
we make efforts to accommodate the researchers; 
which means staying late.  For sorts of cells for 
PCR, we can hold the samples overnight or a few 
days.  On occasion, we pull the plug should 
excessive time be required and a questionable 
yield to be useful for the researcher.
Basically, we provide the service of sorting the 
cells, filter samples (for our own sanity mostly) 
and consult with the researcher for the sort 
parameters.  Purity checks are performed, if the 
researcher is in agreement.  The researcher is 
responsible for all other aspects.
******
>1.  Do you count cells before and after sorts to estimate yields?

Not routinely.	Sometimes users take this on, and we welcome the feedback.

>2.  How do you handle purities on rare event samples?

(I assume that this relates to when few cells are 
collected.)  We generally ask the user if they 
would like to give up some of their precious 
collection to check the purity, but often the 
answer is no so we do not do it.  However, when 
we do, the purity is nearly always very good for what we started with.

>3.  Do you assess viabilities before and after sorting?

Again, not routinely, but we welcome feedback from the users.

>4.  Does your customer have input on flow rate, abort rates, etc?

Yes, but most do not know what to choose.  We do 
our best to make the decisions.

>5.  Do you routinely filter samples in the lab 
>or is it done by the researcher?

We do not routinely filter; however we eyeball 
the sample and if it looks problematic, we will 
do it.	Some of our users know to do it in 
advance.  Otherwise if we are having repeated 
problems, we will filter the remainder of the sample.

>6.  Billing policies on clogging samples, setup

Clogging is usually resolved quickly and we make 
no special billing provisions for it.  Setup is 
billed, but not more than one hour.  We generally 
align our instrument (a Vantage w/ DiVa) and then 
start the clock for billing.  If we do any 
special setup at that point, eg sterilization, we 
bill that time.  Also, we are very generous and 
do not bill the decontamination of the instrument afterward.

>7.  How do you handle timing?	Do you stay late or pull the plug?

Mostly we stay late, but it depends on the 
personal obligations of the operator.  Usually 
our users are considerate enough to know it's an imposition and are grateful.

>If anyone has codified their SOP, I would love 
>to have a copy.  What am I missing?  As you 
>might guess, I am trying to asses what 
>a  sorting facility routinely supplies to their 
>users and what is expected of the researchers.
Sorry I don't have an SOP written, but we ask the 
users to bring 3 things - 1. the sample, of 
course; 2. some extra diluent, in case the sample 
is too concentrated; and 3. some culture media or 
pbs with fcs (or bsa) to collect the sorted cells 
in, or at least to coat collection tubes.  (In 
the case of microplate deposition, we also ask 
the user to bring the plates prepared with media 
in the wells.)	We are usually able to take it 
from there, but as I said, feedback is always 
solicited so we know how we did and what to adjust in the future.

-------------- next part --------------
HTML attachment scrubbed and removed


More information about the Cytometry mailing list