Tuesday, January 10, 2012

What to do with aging equipment: Upgrade or Replace? And a Mini MoFlo XDP Review.

I started writing this post specifically to follow through on a comment I made when talking about our upgrade of our aging MoFlo to the XDP platform (thanks for the reminder Carol).  But then, I started thinking about all the equipment we've held onto and decided to upgrade and asked myself, was it really worth it?  Before I answer that, let's lay out a bit of discussion on the matter, and then I'll finish up with my thoughts on our upgrade to the XDP.

When I think about equipment, I like to put things into 3 categories, namely Cutting Edge, Mid-Cycle, and End-of-life (EOL).  I take all my equipment and shuffle them into these categories and move them around every so often as needed.  This way, I can put things like service contracts, maintenance budgets, capital investment, and upgrades into perspective according to pre-determined criteria.  For example, I'll stick instruments like our 4-laser Fortessa, or ImageStreamX into our Cutting Edge category.  This means they probably won't require a huge maintenance budget since things aren't likely to break yet.  However, they may need more personnel time because the applications performed on them are likely to be complex.  I'll shuffle staff and training resources to those instruments.  Mid-Cycle equipment are things like our 4 year old LSRIIs.  Things are likely to start breaking and so they may eat up some service budget, but they are the workhorses and need to be running full-time.  The applications are probably fairly routine, so they may not require as much custom tech time.  Lastly, our EOL instruments are things like our ancient FACScan and FACSCantos, and it's these instruments on which we need to make decisions.  Depending on the maintenance of these EOL'ed instruments, they may require varying amounts of service and since they may not be as desirable to use as the cutting edge cytometers, you'll need to determine how much money you're willing to invest to keep them limping along.

These EOL instruments can be a pretty decent consumer of budget and may or may not return all of their costs from recharge.  It is with these instruments that we must decide; replace or upgrade (or I guess you could just let them die a slow death).  You'll need to first determine if there is an upgrade path for your instrument.  In the case of the MoFlo, this was a whole-hearted YES, thanks to the good folks at Propel Labs.  Other instruments where this may be a possibility include FACScans and FACSCaliburs, which can be transformed into completely new instruments courtesy of Cytek Development.  If there is not a path to upgrade, then the decision is an easy one.  However, if you're looking into an upgrade, you'll need to weigh the costs against the benefits and definitely compare it to simply purchasing a brand new instrument.  If you're going to shell out a bunch of money on an upgrade, it may make more sense to look into getting a new cytometer.  Sure, you may have to settle with something a bit less powerful, but it'll be nice and shiny and (hopefully) problem-free for a few years.

So, let's put this all into practice with a retrospective look at our decision to upgrade our MoFlo.  We had our MoFlo originally installed in 2000, and approaching 2008, it was definitely showing its age.  Many of the buttons on the "rack" had fallen off, and it seemed like a waste of money to replace entire electronic bays on a rack to simply fix a button.  In addition, parts to fix the MoFlo were somewhat scarce, and it looked like many of the components were approaching their demise.  At the time, options for a new sorter were limited to the FACSAria, the inFlux (both from BD), the Reflection (from iCyt), and the MoFlo XDP (from Beckman Coulter).  All of these instruments easily approached the $0.5Million mark, so buying a brand new sorter without an SIG or a generous donor was pretty much a long shot.  Seeing as our MoFlo was still humming along just fine we decided to look into an upgrade.  The goal going into this thought process was to have a sorter that would handle a lot of the cell line type sorts using GFP or other RFPs and perhaps a few phenotyping experiments.  We did not have the expectation that it would rival our FACSAria and start performing multicolor phenotyping sorts as well as the Aria does.  We also noted that our "GFP" sorts accounted for about 30% of all our sorts and guess what?  We had 3 sorters; a perfect match.  We were able to upgrade our MoFlo for about 1/3rd the cost of a new sorter and get a few more years of life out of it while we waited for the next big thing!

MoFlo-XDP Mini Review:

I can say that the XDP upgrade pretty much met our expectations.  It handles most of our GFP/RFP sorts just fine, and is able to do a few more sorts on markers that are relatively bright.  It by no means can resolve populations as cleanly as our Arias, but it does well enough for many things.  The single best feature of the XDP is zero coincidence aborts.  You may be thinking to yourself, well isn't the Aria marketed as having very low abort rates as well?  It is, but when i say zero aborts, I really mean zero aborts, even when you have 30,000 - 40,000 events going through per second.  The place this comes in handy are rare event sorts at high throughput rate.  We can sort very rare populations and have a really good yield when compared to our Arias. What this really means, however, is that you have absolute control over your yield.  If you need every single cell possible, you can run fast, have confidence that you'll be able to make a sort decision on every single cell, and using a yield sort mode, sort out every single cell.  Sure, it won't be very pure, but at least you have them all and can decide to resort again if you need purity.  Our specification for sorting yield is a 1% population with 70% yield using the purify sort mode (to achieve 98% purity or better).  The max event rate able achieve this on the XDP is about 30,000 eps.  The max rate able to achieve this on our Arias is slightly less (~22,000 eps).  The touch screen is a bit annoying at first, but I've gotten use to it.  The biggest problem with it is the implementation of the slider and up and down arrows.  The slide is way too sensitive, and the up and down arrows are way too slow.   This interface is used for adjusting things like frequency and amplitude and plate voltages.  The new and "improved" Sort Master, dubbed Intellisort, works intermittently for us.  It took a lot of playing around, but we can get it to hold onto a node pretty well these days, but for a while we completely ignored it.  I still think this can be done way better, and apparently Intellisort II delivers, but I'm not going to hold my breath for that one.

So, am I happy with my decision?  Absolutely!  Would I do it again today?  Not too sure. What it boils down to is, I spent a good chunk of change for a sorter that has 4 lasers and about 4 usable detectors at any given time.  The need for the 4 lasers is pretty low, so I could get by with a 2-laser 4-color sorter and be able to do everything I'm currently doing on my 4-laser 10-color MoFlo-XDP.  If I were given the option today, knowing what I planned to use the sorter for, I might check out the possibility of getting a brand new sorter that was stripped down to the basics for cell line transfection sorting.  I'm thinking something like this perhaps might do the trick.