Building a Gel Doc - The secret to perfect gel images
ฝัง
- เผยแพร่เมื่อ 19 ธ.ค. 2018
- When it comes to gel electrophoresis, the only thing more frustrating than running a gel, is trying to see the result afterwards. Our caveman eyeballs just aren't sufficient for picking up the tiny amount of light coming off of a given sample. The solution to this problem? A gel doc. Normally gel imaging systems cost at least 8-12 thousand dollars, but there's no reason to spend that sort of cash. With nothing but a simple wooden box, and filter you can make your own gel doc for less than 100 dollars.
This is one of those tools that when you get into molecular biology you learn to love as it makes your work look more professional, but also makes it a lot easier to see what's going on.
Nerd Thunder Links:
Jeri Ellsworth - / jeriellsworth
David Ishee - / @davidishee
Tech ingredients - / techingredients
____________________________________________________________________
Support the show and future projects:
Patreon: / thethoughtemporium
Nebula: go.nebula.tv/thethoughtemporium
Ko-Fi: ko-fi.com/thoughtemporium
Become a member: / @thethoughtemporium
Store: thethoughtemporium.ca/
______________________________________________________
Our Social Media Pages:
Tiktok: / thethoughtemporium
Instagram: / thethoughtemporium
Facebook: / thethoughtemporium
Twitter: / emporiumthought
Website: thethoughtemporium.com/
_____________________________________________________ - วิทยาศาสตร์และเทคโนโลยี
![Ep.2 เบสไวน์ รวมคลิปฮา SS4 - [ พากย์นรก ]](http://i.ytimg.com/vi/Ihk-LzEbox0/mqdefault.jpg)
Nice work on the gel doc!
For improved images, my strongest recommendation is to improve your gel comb. Your wells are extremely tall. (too long in the axis parallel to the direction of DNA travel) Since agarose gels don't focus the DNA like PAGE-SDS, your band quality and sharpness are adversely affected by tall wells. Not only is your DNA spread out over a wider area (lowering areal fluorescence intensity), that wiggle in the bands is making differentiation of bands difficult.
An interim way to potentially clean up that band wiggle is to flush out your wells with the pipette before loading the DNA. There can often be bits of agarose cruft that sits in the bottom of wells that messes up how the loading gel sits at the bottom of the well.
I had really good luck with DIY well combs by laser cutting some thin plastic. I forget what I used but I just checked Ponoko and they have 0.8mm delrin and 0.5mm PETG, both of which should work fine. A
I love your work
Showing how you don’t need extremely expensive gear to get into genetic engineering, and do it well!
A few things you should try out (from the perspective of a long exposure landscape photographer)
- you did not mention which lens you're using but if it's the kit lens, then you should be able to increase the sharpness a tiny bit by increasing the aperture to f/8 (there's usually a sweet spot and wide open (as you've probably had it) is the worst setting)
- I assume you only used a 10 second exposure because the image became too noisy. Check out your camera settings. There should be an option for "long exposure noise reduction". This will take a second picture (an x second exposure will take 2x seconds) which records sensor noise data. This will remove a lot of noise that is introduced shooting long exposures and this can't be done using software. If your nikon supports that you should be able to take the iso down to 100 by increasing your exposure to a minute or so.
And a few things for others that were not mentioned:
- You don't necessarily need a DSLR. New phones can take long exposures (and even shoot in RAW!).
- Always shoot in RAW never in JPG. This will retain a lot!! more information.
- Don't move your camera and use a timer to take the image. Every tiny bit of movement will blur the image. Even you pressing the shutter button can sometimes ruin the image.
- Lightroom was mentioned briefly but this is the tool that will save you hours if you're doing any kind of photography related task. Photoshop is not needed!
- Always try to keep the ISO as low as possible. Because the camera is stationary you can let the sensor take in as much light as you need.
There's almost no excuse for anyone not to try this and I have to definitely try this out sometime soon. I didn't even know that such things were possible. Thank you!
My first thought was the higher the f ratio the higher the contrast, and I see you have answered this nicely. The low ISO I presume is to keep the noise down on the exposure. I wonder if image stacking could produce better signal to noise as well. Might be an interesting experiment to try. Take a number of lower exposure time pictures and then use image stacking software to combine them. CCDs tend to get noisier the longer the exposure, so by stacking photos you increase the signal to noise ratio.
So a redmi note 7 with gcam be sufficient?
With night sight
Maybe I will shot in a dark room to even further improve the quality
Hey - great job! Btw - that was my article in lab times :)
There have been a couple camera-related comments, but I'd like to add my own:
- noise is a complicated subject in photography, and there are some rare exceptions where shooting above the lowest ISO is desired, but in this case you will almost certainly get a better result by shooting at the lowest ISO you can and lengthening the exposure. If your camera limits exposure length to 30 seconds you may not be able to do this, but try to lower it as much as possible.
- you increase the f number the the 'sweet spot' (dxomark might have info for your lens) but there is a chance this will increase noise more than it will help with sharpness (and sharpness may not be useful in this context anyway).
- disable image stabilization and autofocus, if either are enabled (they don't help in this situation)
- DSLRs can leak small amounts of light through the viewfinder; normally not an issue but can be a problem e.g. if you are doing long exposures or if the sun is behind the camera: try covering it over when shooting
- enable long exposure noise reduction, as another poster mentioned, if your camera has it. If you don't have it built-in you can take your image and then immediately take the camera out, put the lens cap on, maybe even throw it in a sealed bag (in case any light leaking through the lens cap) and take a second identical exposure which you can subtract out in post.
- you can increase the SNR tremendously through "image averaging" techniques. Dial in a good exposure, take 10 of them (more is better). IIRC lightroom can do this itself. If you're using something else: don't do any processing in post besides converting them to 16bit tiffs (no brightness, no sharpening, no lens correction, nothing; only white balance and format convesion). Then you can optionally align them (e.g. hugin comes with "align_image_stack" utility). Then you can use the gmic CLI to align them and do a "median_files" process on all of them. (Make sure to use "-o result_16bits.tif,ushort" on the gmic command line.)
...all that said, it looks like noise isn't really the limiting factor in your images, so much as the contrast between the imaged stain and the glow of the lights in the gel, so maybe it's not necessary to worry about any of the photo nerdery. Might help with your bioluminescence time lapse, though.
Thanks for the awesome video, and thanks for the recommendation on videos
Nice video!!! I should use your video to make a pcr again just like in my time in the university.
Nice built
Thanks for sharing👍😀
Idk why but your videos are really good to draw while watching
Very nice video and attention to detail. The earlier comment about the size of the wells in your electrophoresis unit is true; you could solve this by making a comb out of a bit of 1.5 mm thick teflon sheet: you cut out slots to form a comb such that each tooth of the comb forms a perfect well for your sample . Place in the gel just after your pour it and let it set up. Another idea for the future is fluorescent Western blot photography. I think this system will work for that technique as well.
Thank you!
Thanks for the plug of those channels, really cool stuff. Oh and also,myself, I put my time into IT and programming, so all your bio stuff seems like magic.
The only plasmids I know about are in BioShock.. :)
Nice work. Though ive found working with gelred and gelgreen they are both super sensitive and super bulky dyes. If you are staining before casting and are finding smearing like 7:09 to be a problem try reducing dye to 5 ul (10,000x stock) per 100 ml of agarose gel, 15 ul per 100 ml is a lot of dye. If smearing is still an issue dilute the DNA further or post stain.
Also same volume per well ive found to be fairly crucial so if you are comparing 1 to 4 ul of DNA load 5 ul per well and make the rest up with loading dye.
What would happen if you cracked open the camera and removed the IR filter on it? This is done for astral photography, I'm a Canon guy myself and I know they made a few models specifically without an IR filter over the lens because it was a popular practice. I'm kind of thinking it might let you take like 5 second iso 200 photos which would be really low noise, you could use raw and really average out the brightness and get some serious accuracy.
Im pretty sure the flat top leds actually shine their light out the sides. I wonder if the normal round top ones would work even better. I could be wrong.
Does crystal violet work as a gel stain? It's relatively nontoxic, and supposedly intercalates.
cellphones can do long exposure now for those wondering. my oneplus 6T can do 30 second exposure.
It feels like you could still squeeze even finer precision out of this setup. Why not a longer exposure and lower ISO? If you're limited by the camera's max exposure time, you should get a cheap wired remote and use bulb mode. If you found that longer exposures with lower ISO were actually noisier you might look into whether the camera supports dark frame subtraction for long exposures.
Why would you need to do the correction on the camera? Could you not shoot both the dark frame and the gell image in RAW and do the corection on the computer? I understad you would need to apply the correction manualy every time, as apposed to it being (semi?)automatic on the camera, but I don't see a reason why it HAS to be done on the camera. I also understand that it likely would not work if you shoot in a lossy file format.
Isaac Arthur had a video the other day about 3d printing food. I had an idea from the matrix where you would have a single called organism with Gene's for different flavors and flagella length or something to get different textures. And I thought how could you activate these genes at print time? Maybe an activator that has a light sensitivity to specific bands of light. Maybe start off with and unflavored colony that you shine brown light on to activate chocolate flavorings and then green light on a different sample of the same culture to get mint. ....Walla the start of mint cookies. :) and maybe it's also engineered to be very nutritious. Could this be doable?
I know it's a 1 years video, but i just look at it now, if u read this, have you try some Astrophotography technics ? (take multiple long exposure photo but with not too high ISO and compute them together for extract only the glowing ?)
With all of the experiments and projects you work with, which do you see the most rewarding? What could you see yourself dedicating yourself to for a long period of time if you had to?
May I ask what do you studied and where? I am really interested in biology but the amount of branches and paths available makes me supper dizzy.
You need a uniform distribution of the light! With those LEDs its not so good.
You should try stacking the photos. I wouldnt doubt if you could yield wayyy fainter measurements. Try stacking 5 pics and compare it to 1
these are some cool rothko type album covers haha
whyyy do you need a DSLR. you could use an SLR or mirrorless
It would be interesting to automate the image post processing with a raspberry pi and some programatical DSP.
You can't contact Rosco with a specific spectrum in mind and ask what they have they fits the profile?
Or some Rosco filters equiv for where you are.
you could probably use skiing goggles as a uv/blue filter, right?
Okay, I only know SLR cameras, but ISO 800 seems insanely fast to me. Why? Why would you use a fast sensor? Nothing in the gel should be moving fast enough for this. I'm realizing as I type how little I know about DSLRs... On film, high ISOs are super grainy and awful. But I admit I don't entirely know how any of my SLR knowledge translates to digital. Isn't there a quality loss with a high ISO?
With DSLR long exposures you don't have to worry about reciprocal failure, but you do have to worry about the accumulation of dark current noise. I would guess he found in his testing that doubling the exposure time and dropping the ISO to 400 resulted in more noise.
EDIT: or it could simply be that the camera doesn't support exposures long enough to get away with base ISO
But the ISO is generally independent of the shutter speed tho? Or so I learned. You can have a slow shutter speed with high ISO but I guess then the image would be too bright... Plus with DSLRs these days the image processing tech came a long way and along with the high sensor resolution even an entry level DSLR like the nikon d3300 won't have perceivable graininess up to about ISO 3200...
@@THeMin1000 wow, haven't thought about this for a while. ISO is independent of shutter speed, it's intrinsic to the film's light sensitivity. But I don't know the basis of that, what makes a film more reactive? More silver crystals? bigger? I guess DSLRs need tight small grids of sensors, and they don't change between ISO settings, it's a matter of a long average I guess? Can someone give me a DSLR to play with? I'm very curious. ;)
ISO 800 is more light sensitive, hence making it faster than a low ISO (but not that fast on digital). A higher ISO will be nosier as it picks up more signal noise. However, digital sensors produce more noise the hotter they get (from long exposures). So he chose 800 as it is high enough to use a short exposure (10 sec instead of say 30 or 1 min) but low enough not to introduce to much signal noise.
I don't see why, you can't use a Raspberry Pi and its 8M camera, also the LED could be driven by the RasPi and 'synchronised' in a kind of automated image processing script...
Which UV led do you use ? 365nm, 385nm, 405nm ....
Because that camera has small sensor which would produce pure noise in this setting.
Reminds me of xray graphene lens dopped in.
Good show ,To bad people with bad Taste Can't see It ! Hahaha
"I just happened to have some E. Coli"
*glowing* E. coli!
I have some old romaine lettuce in my fridge, does that count?
Will you do a follow up to kombucha leather video?
Eventually. Got something planned for a month or two from now
@@thethoughtemporium Thanks a lot for responding
I'm surprised you didn't mention Applied Science (Ben Krasnow) this time, when talked about Nerd Thunder. His channel (in my opinion) is also similar to yours.
He'll get a mention. I'll be covering more channels each video till probably the end of January probably
@@thethoughtemporium Oh, I see. I guess, I will have a few more channels to subscribe then.
Your videos are really cool! (First!)
That is one dark box.
Why a box?
Why not an easier to make tube?
Think about it, less corners for light to leak through.
Is it just me or is that tardis blue?
That's really cool!! (Fifth)
Man id sort your gels out before worrying about visualisation
Notice me
I'm scared to know what this guy is doing with DNA considering hes doing this all in his home and all his body modification videos
"all his body modification videos". Of over 100 videos I've published, 4 are body modification. At the moment the big project is this: th-cam.com/video/Fx8TcGrCOSI/w-d-xo.html
But I'll be making and testing a variety of genetic constructs in the new year for various projects