But what will we interface with now? A PIWB MAID? :-)
It is exhilirating to be involved in a project that evolves so rapidly and fluidly! A unique selling point of Crucible (R.I.P)
Apologies for the delay, was on my Blackberry until now so couldn't
see your answers highlighted against my original text. Also my
thoughts on the peristaltic pump idea you suggest!
(Incidentally outstanding Googlemail targetted ads appearing alongside
this email, including the reseller for the MB2, a altitude/pressure
testing company and Fluigent!)
The U2 reference:
Griffin et al Aerobiologia (2008) 24:19-25 - Google it, it's free
For my severe doubts about this, please see my earliest posts on our blog:
Like you, I was suspicious about culturing conditions.
In fact - we'd like to do THREE things:
1 Attempt to collect for culturing - but trying to keep the samples
hypobaric all the way down
2 Carry out a bioassay IN FLIGHT so that nobody could say the result
was caused by lab contamination. Of course, it could still be
contamination from the balloon in flight - see the blog for agonising
about this (self-sterilising balloon anyone?). A suitable bioassay
would be amplification of 16S ribosomal genes with a fluorescent tag.
3 Try to match the two! (i.e see if you could culture something with
the same 16S sequence)
Love to add (4) which would be to try the same microfluidically!
Your later email: A peristaltic pump is indeed a very cool alternative
to a axial fan or a syringe/bellows type arrangement. Normal lab
peristaltic pumps are usually pretty massive affairs though (they have
to be to squish the roller against the tubing really tight). Hmmm....
how about a linear ripple peristaltic pump? Like an inkjet?
I.e a flat piece of tubing being squished in waves along its length to
accomplish the pumping. Might be easier to rig (the SMA actuators I
love pack a lot of muscle for their size and are very simple, but are
linear movement only really)
How fast could we run either type of peristaltic pump though? They are
intrinsically limited to the size of the tubing perhaps? It can't be a
really wide tube or it would be unsquishable (if the walls are thin
then it would quickly tear). If it is small then it takes a lot of
pumping for a given volume. But... I love the linear wave idea, can I
play with it? :-) OOOH that might work for a reel of tubing as well!
(just like you suggested parallelising a normal rotary one)
Looking at that weblink you sent us of the monster peristaltic, you
can apparently have peristaltics with 1/2" tubing!
Another maths experiment...
Let's assume 1cm tygon tubing, which we are more likely to lay our hands on
That has a cross section of about 0.5cm2
1L is 1000cm3
So, you would a 20m length of tubing to pump 1L
(1000/0/5 = 2000cm)
What I was even thinking was - the 16S or culturing reagents could be
INSIDE the tubing already - this would only weigh 1L=1kg in the
example above. But the tubing might be heavier...
(Let's check the axial fan situation further but Fred was adamant they
couldn't shift anything in low pressures)
The balleen idea is a beautiful comparison and I'm floundering how to
think it through biochemically (this is almost an unintended marine
pun! No, to be honest it's an intended one). What would be mixed with
How thin can you make the film? Say it was 0.1mm thick.
Say the axle of the winding barrel for the 1km MAID ribbon I
postulated was 50cm across.
That would be a circumference of about 1.57m
So a 1km ribbon would wrap around the barrel about 637 times
637x0.1=64mm wound thickness
You could indeed either capture bugs on the film and try and grow them
on the ground, OR try to grow them in flight (paradoxically the better
bet perhaps, especially if you have a long flight, since the growth
conditions are most similar to the bugs' natural environment) - but
the problem with those two is you don't know what to grow them on. A
bioassay in the film itself might be best - i.e chamber 1 lyses the
cells, chamber 2 adds reagents, chamber 3 amplifies 16S etc. Mel has
already succeeded with this in the lab - see blog - but not
microfluidically! (Can you do all this in a 0.1mm thick film etc? 1mm?
I've always been fascinated by microfluidics from the outside but know very little of it's parameters!)
Oooh.... here's a lovely idea... I'm going to call it PIWB for Pumped by Inkjet Wave Balleen!
A thicker film with a pump layer made up of microfluidic channels that pump into the culturing or 16S channels - and the pump channels are pumped peristaltic fashion in linear waves, like an inkjet! All integrated in one.
But how do we activate the pump channels? Over to you Gareth!
(In your simpler passive sampler idea, how would you start the epoxy
sealant curing come to that?)
(Reading in the 16S fluorescent results from a PIWB MAID ribbon
(sorry!) is easier - you just reel it in again and scan the ribbon as
it comes back in, like a pianola.)
Night night and best wishes!
I got the basic bioaerosol numbers (I.e 1cfu/L at ground level) from the person (Fred) who builds the MB2 samplier, and I knew him through a science charity, so it was an interesting coincidence!
I think we need to work out some better way of getting the air through the sampler - according to my sums you'd have to pump a 50ml syringe for 60 years!!
Hmmm let's think that through
500ml - 6 years
5L - 0.6 years (219 days)
50L - 22 days, getting interesting
500L - 2.2 days (so a 1 day flight is worth it, the chances of finding something would be 50:50)
That's 500L per return syringe stroke - in say 10s, eqv to a continuous pump pumping 50L/s
The MB2, straight out of the box, is useless b.t.w, because its fan could never pull a vacuum Fred says - it is designed for ground level use whereas we need a pump that pump essentially to vacuum (I.e from 0.01bar, 10mbar, to something less than that, which to all intents and purposes is a lab vacuum).
Fred is quite interested in building us a hotrod version if we can think of a way round it though.
The problem with an axial-based fan design for a pump - what would first spring to mind - is that you can't really pump to a vacuum this way, which is what we are asking (see above)
In fact, electron microscopes etc use plunger-piston type pumps to pull a high vacuum - so not far off a syringe pump after all. You can't spin a fan that fast or that efficiently, but if you withdraw a given volume into a syringe or piston etc then you really have caught something in the barrel, even if it is very tenuous (ie.g 50ml at 0.01bar is eqv to only 0.5ml of sea level air)
Hmm can we cannibalise an EM pump? A surplus one perhaps? But likely to be heavy and high voltage. After all, all we need to pull is 10mBar to 1mBar, not from 1000mBar to zero (which is what the EM pump has to do)
How big a plunger would you need to shift 500L?
500L is 0.5m3
Let us assume a piston with a working stroke of 17cm - 0.17m (this is because this is the useful return stroke of the SMA actuators I have in the lab)
0.5/0.17 = 2.94m2
I.e, the piston has to have a surface area of 2.94m2 for a stroke of 0.17m to have a pumping volume of 0.5m3 (500L)
So say 0.5m radius, I.e a drum 1m across - just about doable, especially if you have a bigger stroke (multiples of 17cm perhaps or a different mechanism altogether) for a smaller radius - OR multiple pistons
Who makes really airtight, sterile pistons? Life support machines perhaps?
The other alternative is to pull a chamber to vacuum on the ground, put it on the balloon, and then suck from ambient air pressure at altitude (e.g about 0.01bar) into the chamber. I have no idea how much a such a chamber would weigh though. Remember it needs to resist 1 bar of ground level air pressure whilst holding vacuum at ground level.
Could you pump it to vacuum in early flight, once the air pressure is less but still "pumpable", say 100mBar?
But is the pump heavier than the chamber?
And so on!!
Any ideas/maths to add my friends?
Can I pump you for ideas? :-)
How much air at this density would a detector 10cm2 square have to travel through to detect one colony?
10cm3 is 1l
Since 1 cfu per 1000l expected:
So, 1000 x 10cm cube=10,000cm
Or 100m air needs to be travelled through to detect 1 cfu at altitude
So 10cfu per km
U2 aircraft have flown similar transects of the high atmosphere.
A 1000km U2 flight could therefore theoretically detect about 10000cfu.
But only a single colony was detected during such a flight in real life using a similarly sized detector, suggesting that at the culturing conditions used in the experiment is question, biodensity is about 10000 less than at ground level, even correcting for altitude (and actually U2s fly slightly lower and atmospheric pressure is somewhat higher than 0.01bar). And that's assuming the single colony seen wasn't a ground contaminant.
If it was 10,000 less biodense then on the ground, then this means (allowing for altitude) not 1 cfu per 1000l but 1cfu per 10 million l !
But balloons fly far less far. If one opened a sampling port at final altitude, this suggests you'd need to allow the balloon to drift 1000km - OR pump 10 million L through your sampler.
Say you have a 50ml syringe pipetting up and down. 20 strokes is 1L
You would therefore need approximately 200 million strokes to detect one cfu. If each return stroke took 10s, this is 2 billion seconds
This is about 60 years!!
A long duration drifting balloon, flying thousands of kilometres, therefore seems more feasible.
One must also however be aware of sample port size.
The above calculations are for a 100cm2 detector (eg 10x10cm)
If your sampler had only a 1cm2 aperture, this would obviously be a hundred times less. Then the balloon would have to drift for 100,000km to detect 1cfu!
Could a large volume pump be developed, for instance using bellows?
Alternatively, could the sampler area be massively increased?
For instance, imagine a Massive Array of Inexpensive Detectors, MAID - a ribbon 10cm by 1km perhaps. That's 10,000 times the surface area of the original 100cm2 detector concept.
Then you would only have to drift for 100m to detect one cfu!
Of course there may be organisms that are hard to culture but which can still be detected by RT-PCR for instance. But these calculations certainly set an upper pessimistic boundary of what effort is needed to biodetect at altitude.
How big a cavity is 10 million litres? Could you perhaps sample this big a volume using the interior of the balloon?
10 million litres is 10,000 cubic metres
4/3 pi (r)cubed
So 30000 = 4 pi (rA)cubed
So 7500 = pi (r)cubed
So 24000 = (r)cubed
So the cube root or 24000 = r in metres
This is a radius of about 29m
So, a balloon 58m across has a volume of 10 million litres
Is this too big? How much does the canopy weigh? I have no idea.
One could in theory fly the canopy with helium to altitude, then rapidly deflate it and fill it with ambient air. If it was fully sealed, the canopy would then rapidly descend and collapse upon itself since the exterior air would be getting much denser. You would be left with a crumpled canopy with a volume, at ground level, of about 100,000l - a precious sample of high-altitude air of a sufficient amount to possibly replicate the U2 experiments.
Please feel free to pick this apart/add references! I can just about believe the U2 results from this (so much air sampled along the transect) but now have doubts about any bioprospecting results claimed from high altitude balloons so far, unless they drifted for long distances (1000s km) before recovery. Both of these types of experiments involved culturing on the ground, so as always ground contamination, or contamination of the balloon before flight or by aerophiles in the troposphere, remains a possibility.
Our efforts to carry out experiments at altitude negate this BUT how do we sample enough air? Must we fly on a long duration balloon now?
One cfu literally means one single bacterium landed on the detector and was able to divide and grow into a visible colony. In turn this would literally mean only single copies of key genes to be detected by us by RT-PCR etc. In short, it doesn't matter how sensitive your assay is if the odd bacteria you are chasing in millions of litres doesn't happen upon your detector - otherwise there is nothing to detect!
Long duration? MAIDs? (Microfluidic?) Collapsed canopies? Any thoughts gratefully received!