Cheap capillary electrophoresis

[Harristotle]

Hi to all.
 
I shared this with my local group (DIYPerthBioHackers), but I thought it might be of more general interest.

https://www.youtube.com/watch?v=jSwYqz_R9n8&feature=youtu.be

I'm a protein man myself, with relatively little interest in that other conformationally boring thing, whose acronym could stand for Dont kNow Anything ! However, the world has moved on, and I am interested in dragging myself out of the Jurassic.

Capillary electrophoresis seems like such a good fit for DIY bio: No gel media, cheap sybase dyes, good possibility for optical detection. I thought I would have a go, and this video is my report of my first efforts.

I note that Raspberry Pi 3 will run java and ImageJ quite well. I also note that a close magnifier can be made out of a 40x objective lens and a pi camera. A script to grab a frame per second, and then count coloured pixels using imageJ would give you a count of stained DNA, or even fluorescently labelled versions of the same, if your home made pcr was good enough to try sequencing.

The power supply shown here is sub $5 (or was when I bought it), glass capillaries are easy enough to pull, and silanisation is possible to minimise DNA sticking to glass if this is an issue.

As I said, I D(on't) kN(ow) A(nything) about the big boring molecule and its chemistry, except to note in passing that it is no longer referred to as sodium thymonucleate, which was on the only bottle of the stuff I ever owned!  I would appreciate any advice as to the feasibility of my idea.

I guess if my DNA ideas are crap, I could always try and re-purpose the design for a little electroosmotic pump and a little HPLC - now I'm getting into more familiar territory !

Cheers,
Harristotle

[Gordana Ostojic]

 

I can't see from the video the other end of the tube and if that is in liquid. Anyway I did use similar thing (it's somewhere on this board). It turned out that optical detection is hard (material issues with UV, lamps are expensive, power is demanding...) so I found out most people work with dielectric detection. There are number of papers for portable, cheap electrophoresis based on this (if you need them I can dig them out).  For DNA you do need gel-like sieving medium like some big polymer along with the buffer again there are papers for this. 

[Nathan McCorkle]

On Mon, Apr 10, 2017 at 10:59 AM, Gordana Ostojic
<gordana....@gmail.com> wrote:
> Anyway I did use similar thing (it's somewhere on this board).

I was reading your post a few weekends ago, and ordered carbon tape
and some platinum wire... along with a laser cutter so I can make some
fixtures (or possibly lithography masks, so I can etch some glass).

Hoping to experiment with a DNA ladder in a month or so.

[John Griessen]

On 04/10/2017 01:23 AM, Harristotle wrote:
> Capillary electrophoresis seems like such a good fit for DIY bio: No gel media, cheap sybase dyes, good possibility for optical
> detection. I thought I would have a go, and this video is my report of my first efforts.

What is the capillary initially filled with, a conducting buffer solution? No gel needed, just the narrow capillary path?

[Gordana Ostojic]

For DNA, it needs buffer soluble polymer as a sieve (pvp for example), matrix flows but slows everything. Companies use PA gel inside for high resolution throughput. That one is fixed.  

[ukitel]

Hi,

I would be very interested in those papers!

We were discussing about capillary electrophoresis just last week and we wanted to try out something...

On Monday, 10 April 2017 19:59:50 UTC+2, Gordana Ostojic wrote:

[Gordana Ostojic]

Here is one

Good luck!

[John Griessen]

On 05/03/2017 08:52 AM, Gordana Ostojic wrote:
> Here is one

.
.
.

Hello Gordana,
You said about harristotle1's video: https://www.youtube.com/watch?v=jSwYqz_R9n8&feature=youtu.be

> I can't see from the video the other end of the tube and if that is in liquid. Anyway I did use similar thing (it's somewhere on
> this board). It turned out that optical detection is hard (material issues with UV, lamps are expensive, power is demanding...) so
> I found out most people work with dielectric detection.

Are you sure optical detection is still "hard" today? How did you do the dielectric detection? Measure capacitance of the
capillary circuit?

[Gordana Ostojic]

Optical is more expensive for sure. When I looked at it, it would cost more than $100 for each, diode and detector. UV lens is also a problem (glass needs to be UV transparent) but I found a neat paper how to make one from pdms. I bought an old spectrometer and used its lamp and detector with this pdms lens. Got the signal when I tested concentrated dye, but it was very weak. Later I figured that I should have removed the capillary coating. So sensitivity may be better for optical but cost-wise, alignment wise (aligning UV optical path is a dangerous and hard thing) , stability is better for capacitance sensing. Capacitance also has its problems, I later figured that I stray currents from HV can mess up the circuit so better electrode geometry and enclosure would help. Unfortunately, I have to do other things now and the electrophoresis will have to wait. 

[Harristotle]

Hi Gordana,
what I had in mind for optical detection was a simple USB microscope focused onto the capillary. You could then take a video of it (amcap), and then pick the video up with ImageJ and plot color vs time/frame number.

My ImageJ skills are weak - I knew it as NIHImage, and I never had my own Mac to run it on, so I have not yet figured how to do this. I know from the examples that it is possible to step through the video (avi) file and treat the images sequentially. I am short of time, so only playing with this a little.

Possible cheap detectors for this:

1) a uv source shining through the capillary onto yellow highlighter drawn on paper behind - monitor diminishing of the fluorescence when an aromatic floats past.

2) some variation of sybase green/ DNA ladder


I am quite impressed with the magnification/focusing possibilities of those usb cameras. It is not near spectrophotometer level, but it is very cheap, and looks quite robust. When I get it all going, I'll look at linearity.

My current update is that capillaries with silicone oil run well with pillbox red and 50mM NaHCO3 as running buffer (pH 7.8, supposedly, but not measured).

In answer to your other question, both ends of the capillary were in fluid.

Cheers,
H.

[Gordana Ostojic]

Did I understood correctly? You want to use florescence as a detection instead of typical absorption and you can do that with labeled DNA. It might work in that case. Like I said UV is a problem since you need special optics components. If you just need UV excitation you just need to make sure that the microscope slide that the capillary is fixed on is UV transparent (and maybe remove the capillary coating). I like the idea of using microscope optics but florescence will be weak (small sample amount and florescence is weak in general) but adding all the pixels will help. One problem that could arise is that DNA ladder needs to have polymer passed with them. These would also absorb excitation wavelength and considering higher mass/molarity they would probably reduce your signal a lot. Another thing is heating, florescence would increase heat and that is not good for capillary stuff. I am curious how it will turn out.

[Gordana Ostojic]

Actually considering that labeled DNA in gel is visually florescent it might be ok, depends on exact polymer and wavelengths used. I would just test it in bulk first before going through capillary, just in case. Good luck!