h a l f b a k e r y
The embarrassing drunkard uncle of invention.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
or get an account
A long time ago, centuries, we all know the stories of how
didnt have the hygiene we today have, because germs
been discovered, because we didnt have the apparatus to
see them. Operations would turn out badly because of
infection, and nobody knew why, it was just one of those
Then optics allowed things both small and near to be seen.
And also large and far away, but never mind about that for
now. Microbes could be seen to exist, and their effects
be studied and predicted. Soon, the beer and wine that
people had been brewing for centuries could now be
in reality because yeast was finally understood. Operations
would have a higher success rate because doctors would
their hands beforehand, now that germs were understood
exist even if you couldnt see them with the naked eye.
The public opinion soon swung around to fear the unseen in
totally new way there really were germs around, too
to see. There was even a widespread fly scare [look,
down there] which
increased the general fear of every surface around.
Today, despite our modern enlightenment thinking, we have
just as much a misunderstood fear of the unseen. We think
germs are everywhere, and they are all harmful, and we
never allow food to come near surfaces or planes that may
have come near carriers that may have come near germs.
we assume the floor is comprised entirely of germs, and
if something goes near the floor on its way to our mouth, it
will kill us.
We simply dont know what may be there, unless we carry
around a microscope, and with the exception of a large
portion of people on the half bakery, probably nobody
actually does carry a microscope around.
I propose a simple invention. A germ sensor that can be
integrated into a phone, so that people can indeed carry a
microscope around. Not just a microscope, but an
one, that can tell you not what germs there are (who cares)
but what the real danger is. Otherwise, were back to
guesswork and assumption and a (now enlightened)
fear of everything. I would imagine the germ sensor might
mostly optical imaging but could alternatively be based on
acoustic snapshots, or even a combination, to identify such
The great fly scare
[Ian Tindale, Apr 08 2016]
The hygiene hypothesis
There likely is such a thing as too germ-free [CraigD, Apr 14 2016]
||The higher-resolution a phone camera, the more that
"zooming" an image is possible, resulting in the effect of
magnification. So, all the phone really needs is a high-
enough resolution camera to directly image germs.
||This is a great idea, except it's all wrong. Microscopy
is about the worst imaginable tool for detecting
bacteria, let alone identifying them.
||What's needed is a little portable PCR machine and a
few other gubbins. These things exist already, but
are not really aimed at the average consumer.
||You could replace it with a box with a red LED with the words, "if LED is lit, bacteria are near this device" Always on.
||//The higher-resolution a phone camera, the more that
"zooming" an image is possible, resulting in the effect of
magnification. So, all the phone really needs is a high-
enough resolution camera to directly image germs.//
||//Microscopy is about the worst imaginable tool for
||The problems with imaging bacteria are multiple and
insurmountable. If you google "Bacteria" and look at the
image search results, you will see diagrams, cartoons and
false-colorized images that look all 3D and cool. The
problem with those is they were not made using a light
microscope they used electrons. There is a total absence
of cool light microscopy images like you may see with
||The key to problem 1 is in the name. Microscope. You can
only effectively image down to the micrometre range.
You can't just keep zooming in. Eventually, you run into
the Rayleigh limit. Light is too big. The smallest visible
light photons have a wavelength around 400nm, or 0.4
micrometers for simplicity. Your effective resolution, is
governed by the size of light, you can tell two points
apart if they are separated by more than about half the
wavelength of the light you are using. So about 0.2
micrometers for violet, and about 0.4 for red.
||Your average bacterium, if there is such a thing, is about
1-2 micrometres long. So, if you really push it, you may
be able to get a 3x9 pixel image of the bacterium you
want to look at. To get this image, you have to do a lot
of things. If you think that getting a non-wobbly image of
your dog at the end of the garden is hard... not wobbling
by a single nanometre is VERY much more difficult. You
also need a pretty remarkable lens. It needs huge
curvature to get the magnification you need, that
curvature has to be accurate to beyond the nanometre
scale, how they do it is very much in the realm of
industrial secrets. Despite the huge curvature needed,
the side of the lens facing the sample must be VERY flat.
Then, because light travelling through different
substances bends differently depending on its color, you
need an additional corrective lens behind the first. Then,
because different parts of the image have passed through
different thicknesses of lens, you need to correct for
that. Why do you need the lens? well, because the
amount of light coming off one bacterium is basically 0, in
regular room light you're looking at 1-100 photons per
minute. You need a GREAT lens to get about 15% of them.
It's ok though, those lenses are very much baked, I have a
used one that i can let you have at the bargain price of
$45k. So you get the image, what would it look like?
||Here's problem 2. Cells, of pretty much every kind, are
bags of water. There's simply not enough optically
interesting stuff in a 1 micrometer thick bag of water to
generate enough contrast, worse, remember the number
of photons? If you just set up a microscope to look at
hulking great mammalian cells in white light (light is
transmitted, through the sample into the lens), you MAY
see the nucleus, a few vesicles and with a trained eye
some of the outline. Microscopy has relied on ways of
generating contrast. So, you flood a sample with a dye
that stains DNA. Your 3x9 pixel image may have a couple
of pixels darker than the other. That's not the best
||Now, can someone tell me why an electron microscope is
called an electron microscope? when it's exclusively used
to image at the nanometre range. I'm off to vandalize a
||As you hint, optical techniques are probably not
useful here. Unless there we forgo the prospect of
capturing and identifying one single small organism,
and instead go for populations. I suspect the acoustic
footprint or fingerprint or signature would enable
detection and differentiation of enough of a
population to have an effect on us, without
necessarily being able to resolve a single one.
||The main aim of this is to demonstrate that there is /
might be / probably isnt a situation where actual
germs are actually there, over there, on that area of
the surface, and it will / might / wont harm us if we
look at it / touch it and then touch our face / shove
it up our arse.
||[bs0u0155], germs were discovered long before electron
microscopes existed. That means light-focusing
microscopes were good enough to at least see them (which
is what I thought could be done with a high-enough
resolution CCD camera), even if identifying them beyond
basic shape takes an e-scope. I don't challenge your
statement that good-resolution imaging of germs won't be
possible with a mere CCD camera. But simply noticing they
||//I suspect the acoustic footprint or fingerprint //
||Yes, the acoustic footprint would do it. Most
species of bacteria (especially the gram-negatives)
produce distinct ultrasonic frequencies as the
bonds forming their cell walls are broken and
remade. A group in Japan has shown that the
frequency and timing of these sounds can be used
to identify bacteria down to the genus and
sometimes down to the species.
||So you could probably do this with a phone if it
had a very good microphone. Or at least you could
if what I had written above was not complete
bollocks, which of course it is.
||The way to do this is for the app to guess, based on image / color analysis, what type of germs are likely there. This is mostly to keep things consistent so a given surface (like your hand) is not proposed to have some different germ each time you look. Bacteria of a given type all look the same and so one would then use stock footage.
|| There would be an option to click through and see what the likely origin of a particular germ is: dog feces, human feces, cat feces, mucus from infected animal glands, moldy meat, etc. A further click through would have images of persons suffering from the diseases caused by that particular germ and an opportunity to purchase specific countermeasures.
||//good-resolution imaging of germs won't be possible
with a mere CCD camera.//
||The camera isn't the problem. In fact we have a snazzy
new super resolution microscope which pulls a few tricks
to get down to the 20 nm kind of resolution. It has a very
fast and sensitive camera, but we only use 1/4 of the
sensor at a time, so only 512x512 pixels. Its trivial to
expand the image to the whole sensor, but all you're
doing is spreading the same light thinner. You gain no
more information, because down at the pointy end there
is no more information.
||//even if identifying them beyond basic shape takes an e-
||Ah, but no! Those contrast tricks we pull get pretty fancy.
Initially, people found that some bacteria stain purple,
some don't. Now we can do all sorts of things with
fluorescent antibodies and so on. Way easier to just do
the DNA stuff though.
||//Bacteria of a given type all look the same//
||Indeed they very do. For example, Yersinia pestis
looks identical to Yersinia ruckeri. The only
difference is that one is bubonic plague whereas
another only bothers some species of fish.
||Or take E. coli (where are italics when one needs
them??). Your gut is absolutely swarming with
gazillions of them. But E. coli O157:H7 has a fair
chance of killing you stone dead.
||Still, this is a brilliant idea if we can get around
the two problems of light microscopy being useless
and appearance being irrelevant.
||Easiest thing to detect is moisture, in my opinion. An old saw suggests that all fungi need to grow is dark and damp, and my understanding of other microorganisms suggests the same conditions apply to them.
||1. A phone would probably detect its own presence first and most strongly.
||2. An accurate device to detect ribosomes would at least tell you if an object has been exposed to a living being, to wit, and is contaminated.
||"An accurate device to detect ribosomes" would be
rather elaborate. Better to simply test for bacterial
rDNA, which is abundant and very tractable to
||That's okay, Maxwell, but I think easier to detect if people have been there in order to guarantee that germs are there.
||Ribosomes are a common contaminant to both domains
||How about an app that just screams "Germs ! Germs everywhere ! Trillions of them ! Everywhere you look, germs ! You're all going to die ! "
||That would neatly cover the known facts.
||//Ribosomes are a common contaminant to both
||I think you mean rDNA. Ribosomes themselves would
be pointlessly difficult to detect (or tell me how you
planned to do it). rRNA is also abundant and fairly
easy to detect, but any sane person would opt for
||This is the halfbakery ....
||Disclaimer: nothing I say should be construed to be my effort to prove my sanity. :0p
||I understand you're a microbiologist, Max, so freely point out my errors; but, how I'd proceed is use a molecular sieve followed by rna differentiation. Finding a ratio of protein to bacterial rna suggests a path to characterizing any microorganisms; maybe if sensitive enough one could differentiate riboswitches from Rho proteins, for example.
||But just being aware of the presence of bacteria doesn't tell you anything. 90% of the cells in the human body are bacteria cells rather than human cells and, mostly, they're nice, beneficial bacteria.
||Thats just it. Detecting actual microbes is not what
this peripheral or transducer should spend its time
doing. Discriminating actual danger (from microbes)
||The issue is that there are people and theres the
other sort of people. One sort will lay a sandwich
down on any surface, pick it up and eat it, and
whatever drops out onto the floor goes back into the
sandwich, if it doesnt go into the mouth first along
the way. The other sort would be convinced theres
dangerous germs just there on that surface,
immediately under the sandwich, and now the germs
have walked over onto the sandwich. The floor
contains germs just waiting. The outer bags of
anything that was bought in a shop, then went
outside to get to be brought home, are now covered
in germs, including from all the other people at the
shop. The outer bags must not touch each other or
more inner food. Letting bare food touch one of
those outer bags is lethal.
||My point is, where are these alleged germs. Obviously
there are them, and they can cause harm, but really,
are they absolutely everywhere all the time no
matter what? In which case we may as well give up
and take no special measures. Or can I point out that
it was okay to put the biscuit down on that surface,
because I know for a fact that there were no harmful
microbes at that particular spot at that particular
time, but had I put it down 3.5cm over to the left,
well, thats a different matter, there was a harmful
colony of bacteria there at that time just as it
||We need the proof. Proof against mindless paranoid
fear of the unseen. Its as bad as fearing the wrath of
invisible gods all over again, you see. Technology
should be able to define and discriminate and allay,
||But this approach will create an environment which applies an evolutionary pressure to bacteria to select for bacteria which are dangerous but closely resemble 'good' bacteria.
||» Which in turn drives evolutionary pressure toward
Android / iOS developers of IoT stuff to evolve
cleverer discrimination and clue finding means,
probably involving google. Obviously humans wont
keep up so our first step would be to develop
that can, using hyper-technology, do the research,
development, prototyping, assembling, mineral
mining, logistics, making, design improvement,
and banking, all by themselves, in a perpetual arms
race to the bottom of the peak of development of
the war between harmful invisible germs, and
beneficial invisible manufacturers.
||This is, of course, assuming, as you imply, that,
harmful bacteria (or other sorts of camels) have an
advantageous payoff from harming humans in the
first place. Im not sure theres an equilibrium there.
Id posit that the thriving or otherwise of colonies of
harmful bacteria versus beneficial bacteria is not
directly mechanically linked to the well being of
passing nearby humans.
||If there were a direct causal relationship between
the more the betterment of a colony of harmful
bacteria linked to the more the worsenment of a
casual passer-by, then maybe.
||//how I'd proceed is use a molecular sieve
followed by rna differentiation. Finding a ratio of
protein to bacterial rna suggests a path to
characterizing any microorganisms; maybe if
sensitive enough one could differentiate
riboswitches from Rho proteins, for example.//
||Horribly, horribly difficult. The world is full of
protein (and somewhat less full of RNA, since it's
||Ribosomal DNA is present in multiple copies per
bacterium; is relatively stable and easy to detect;
and can easily be analysed in detail to distinguish
bacterial species if needs be. Plenty of modern
diagnostics rely on this. (In fact, I'm currently
working on a fast assay for neonatal meningitis,
which has to be able to detect 1-2 individual
bacteria in a small sample of CSF; it uses bacterial
rDNA for exactly this reason.)
||:) nice stuff, slow and tricky if one goes about decreasing chemical activity, elution, PCR, and all that.
||Do you take a different approach, like spectroscopy or cross linking antibodies?
||You mean me, for the meningitis stuff? It relies on a
two-stage PCR (which can detect single DNA molecules
pretty well), targetted at the conserved regions of
bacterial ribosomal DNA. The clever bit is the way we
eliminate contaminants in the system - and believe
me, there is bacterial DNA in **everything**, even
||I should add that my test is only aimed at giving a fast
yes/no for bacteria - in a neonate, the most important
thing is to know if there are *any* bacteria in the
spinal fluid: you can worry about what sort they are
later. If there are definitely no bacteria, you can send
the mother and baby home; otherwise you have to
keep the baby in for 2-4 days on precautionary broad-
||Doesn't taking a CSF sample risk infecting said neonate?
||Risk of infection vs. risk of misdiagnosed/untreated meningitis ?
||//Doesn't taking a CSF sample risk infecting said
neonate?// Not if it's
done properly - apparently the risk of infection is
negligible. Also, skin-
borne bacteria don't appear in the CSF samples,
which implies that they
don't get carried in on the needle. I've not seen
them taking the
samples, but it's considered a low enough risk that
it's done routinely for
babies that seem off-colour or have a high chance
of being infected.
||At the moment, the CSF samples are cultured for
48 hours to see if any
bugs grow, which means precautionary antibiotics
and a hospital stay in the meantime. By
the time the culture results come back, the kid's
probably OK anyway.
And only a small fraction of "positive" samples are
either because the bugs just don't like growing in
culture, or because the
mother has been dosed up on antibiotics before
the birth as a
preventative, and any bacteria in the kid's CSF are
||The main aim of my test is to be able to send
home the definitely-
uninfected babies immediately; so it has to have a
very low false-
negative rate but can afford to have a modest
false-positive rate (since
the procedure then will be the same as if there
was no test - i.e. treat).
||Clinics here have been using fast MRSA and Influenza A screens here for a while - just to throw ab/ics at bugs with a degree of sensitivity. Getting babies out of the hospital before they contract any MDR cooties is paramount. Cool work, Maxwell!
||//fast MRSA and Influenza A screens//
||Yep - most testing now is moving from culture over
to PCR and/or antibody-type screens. The
problem in my case is that (a) there
may be, literally, only one or two bacteria in a
relatively large (200µl) sample and (b) because
we're testing for *any* bacterium (since many
different ones can cause neonatal meningitis),
we're vulnerable to contaminants by *any*
bacterial DNA. Even ultra-pure, sterile reagents
have significant amounts of bacterial DNA on or in
||//one or two bacteria in a relatively large (200µl)
||Why such a small sample? I'm not sure how much is a lot
in neonatal cases, but instead of taking a small volume,
why not ignore all the irrelevant salty water? You could
use a standard needle/syringe with a tiny intermediate
bit of plumbing, pull 500ul, push back through a one way
valve with a bacteria-binding filter. Rinse and repeat for
a couple of ml. Also, why are DNA-free solutions difficult?
I have a total nightmare of a time keeping Ca2+ in the
nM, but EVERYTHING is contaminated with Ca2+ (when
you think about it, the Himalaya is just a very large Ca2+
contamination issue) and it's essentially indestructible.
||What about ethyl methane sulfinate as an alkylating agent and potent mutagen; or a repressor protein that would block the operating cells?
||Well, it wouldn't obstruct replication but it does give rise to an
error in replication so that the newly formed DNA strand carries
a mutation and you've got a virus.
||So I shouldnt put my biscuits there?
||Only if you're happy for them to burn twice as bright, but half as long.
||//Why such a small sample?// The standard
collection is 5 drops (each about 30-50µl) into the
collection vial, and usually 2 vials.
||I agree, you could re-inject the fluid via a filter to
recover the bacteria, but then it becomes a rather
different procedure, as far as the doctors are
concerned. It's very difficult to get clinicians to
change what they do, because they are at the
sharp end of things, so to speak. There may also
be some difference in regulation (for instance, a
re-injection system may come under a different
category of "medical device"). It would be useful,
but it's got to be proven useful and has to be
tested to hell and back first.
||Maybe with the right nanotech beads a microdialysis catheter could provide both a few millimeters surface area for an RNA footprint and draw essentially no CSF. I don't know how, but thinking it out suggests to me optical fiber and fluorescent "wells" lit by, I don't know ... photon kinetics or chemical kinases?
||How about Raman spectroscopy?
||nice, electron nanoscopy.
||The more bacteria, the higher chance of some bad. But still, it has to be on negative volatile molecular species, the more common the better. I have a cellphone that reacts nicely to too much meths, especially when I'm cleaning splashed paint.
||That is, be mobile germ sensors... and roll in them.
||Better than children, which are self-propelled disease incubators.
||How else is the generation barrier broken?