Sponsored by Blinkist. Get 25% off Blinkist premium and enjoy 2 memberships for the price of 1! Start your 7-day free trial by clicking on this link: www.blinkist.com/asianometry
What do you think of the size of economy, US vs China? Whose economy is bigger? I was surprised to find out that the US economy is probably less than half the size of Chinese economy. The sales of luxury cars in 2021: Benz: 758,863 sold in China, 276,102 in the US BMW: 846,237 sold in China, 336,644 in the US Audi: 701,289 sold in China, 196,038 in the US Porsche: 95,671 sold in China, 70,025 in the US EV: 2.8 million in China, 0.4 million in the US I looked into further, and there are lot of other evidences that prove China is bigger, has lot more money and resources than the US does.
One simple nitpick at the beginning: at the moment, next generation sequencing such as illumina is no longer called NGS since they’re … current generation. They’re called high throughput sequencing since 2018. Next gen sequencing are called long read sequencing, to avoid this exact problem SNPs are often not spell out but pronounced as “snip” As for 3rd gen sequencing, or long read sequencing, the field has pretty much settled on either pacbio or ONT. Hot topic right now is actually hybrid sequencing, where long read and short read are combined to improve both coverage and accuracy of read, as well as remove the bias introduced by the ref genome. We need to dive a bit deeper into how genome assembly works here, but in short we often align reads to a reference genome (from the human genome project for example). Now we can use the long read from Nanopore or PacBio as the reference, and assemble the short reads to this ref. A point about the consumables for illumia: the reason we need way more nucleotides than present in the genome is because we underwent the amplification process. If we have an amplification factor of only 10 for all basepair, that bring number of nucleotides to the 10s of billions range. Account for waste and inefficiency, the consumables rate is reasonable.
What about the sci-fi sequencing where you use nano-machines and SEM to actually read tiny bits of DNA (without PCR amplification)? How's that going? A decade or two ago that was the "next big thing". Haven't kept up with the field since then.
@@raylopez99 that sound like nanopore technology (ONT). It’s going strong! They have a version (MinION) that about the size of an external SSD. Super neat for field works and environmental science folk. As a computational scientist, id say they have their problem with bias and lower accuracy than amplification methods, but they can be supplemented with short reads to build a robust and cheap sequencing
Everyone I work with simply refers to Illumina as NGS. Perhaps as consumers rather than developers of sequencing technology we’re too disconnected or lazy to bother making a distinction. Though it does speak to Illumina’s dominance in short reads. I agree that ‘next generation’ has been a comical misnomer for a decade at this point.
@@VuLamDang ONT's long read quality is better than you think. I routinely get quality scores comparable to illumina, every year they make huge strides on that, and now they have the advantage of being able to detect modified bases much better than any other method. If your experiments need long reads and nucleotide modifications, then ONT is really the only way to go.
Hi Jon! I am a sales specialist for Illumina Taiwan, delighted to see your video on Illumina! I have been following your channel for a while, really enjoyed the story about Taiwan’s camphor production history!
Hi Thomas, greetings from JenOptik ;-) We'll send a bunch of new MNSQ & NSQ readheads out of the manufacturing to singapore tomorrow, make shure you'll sell them well. ;-)
Hope you enjoy those exorbitant stock options while the company is in severe distress!! I hope the owners stop paying so those who are corrupt leave or get FIRED!!!
I appreciate how you described the process of sequencing and the history that led to instrumentation. I was trained on Illumina HiSeq 2000s, MiSeq, MiniSeq, and NovaSeq (briefly) when I was a molecular technologist. They're easy to use, and technically interesting on how they operate. The modular design of their more recent line of sequencers helped with cost and troubleshooting, and when I was a supervisor (I'm in Pharma testing now) this helped decrease the pricing for patients. Easily could design testing under $300 for most panels, and cheaper with higher volume.
Funnily, the HiSeq 2000 used readily available microscopes as the core body, they wrapped motorised parts around it. A lot of components are off-the-shelf systems integrated together. Despite the model being phased out and recycled, many parts are useful for optical builds to this very day. I pulled a couple apart and almost all the components have been reused at our institution. Recycling, done right!
@@timmata8143 Its ok. Its just some additional money you have to spend. But I remember the reaction of one of our colleagues in the IT department, when I requested some computing resources. He proudly offered me a VM with 16(!) Threads and up to 16 GB RAM. So I said: You did nothing wrong, everything is OK, but I need to talk to your boss. ;-) Right now we run two servers with a total of 1.75 TB RAM and 256 Threads. Do not get me started on storage. Sore topic.
Wow. I cannot imagine how much time he puts in to do the research, find the photos, edit the video, recording, and organizing, to create this video. Good job.
@@orangestapler8729 It really is master level. You can't show me a single teacher or professor researching and making a material like this for their studies let alone the delivery.
It's missing the Polymerase syntesis, that allows to indefinetly grow the gen sample (the "amplify" of the gen sample you talk about, is done with it), discovered by Margarita Salas, and that allowed the start of the human genome project 1 year after
@asianometry, I just want to say 'THANK YOU' for the work you do on your videos. I always learn a lot watching them, and really enjoy your straightforward, thorough coverage of whatever field you're looking at. And your presentation is great - not going off into tech-speak, but not talking down to the viewer either. I learned a lot about gene sequencing from this video, and have a bunch of things I now want to go look up to learn more. Well done!
I really love how well detailed you've explained these concepts, for a gu your background, people find it hard to grasp. Great content . Molecular biologist here 🙂
I do a lot of DNA sequencing for my work, and really enjoyed your recap of the history of sequencing technology. A fun potential future video topic in this area: will nanopore sequencing, and Oxford Nanopore specifically, supplant Illumina and optical sequencing-by-synthesis over the coming decade? A custom ASIC is at the heart of Oxford Nanopore’s flow cells, their CTO originally worked at Solexa, and their products seem (to me) to be right on the cusp of surpassing Illumina on key sequencing quality and quantity metrics.
At the time when I used them, sequencing of homopolymers was a bit of an issue with the Oxford Nanopore MinION. Similar to the iontorrent. Back then the algorithms would sometimes miscount the number of nucleotides in this type of sequences.
@@johnsmith1926 I'm curious as well how conventional sequencing avoids misinterpreting long stretches of one nucleotide. Wouldn't synthesis proceed right through that sequence and the fluorescence not be able to distinguish it's length?
@@samrusoff from my understanding, the florescent molecule prevents further nucleotides from binding to the RNA strand being synthesised, meaning that for any given cycle, only one nucleotide can join to the end of a strand. The remaining reagent is cleared away, image of the clusters taken, then a buffer is pumped through the flowcell, destroying the florescent molecule before another set of reagents is pumped through to begin the next cycle.
@@johnsmith1926 Homopolymer basecalling remains a challenge, but not an insurmountable one. ONT has switched to nanopores with longer channels, so they read a longer stretch of nucleotides and can accurately basecall up to 10-20 bp homopolymers. And in the last year they’ve tweaked their flow cells so that for a large fraction (up to >50%) of the DNA molecules they read, they thread both strands of the duplex through; and by basecalling with information from both strands of the duplex they can get to Q30/99.9% accurate single molecule reads. This means that for applications where exact basecalling of long homopolymers matters, it should be possible to use a ‘polishing with dirt’ strategy, where you split and barcode a sample into two: one half that you prep and sequence regularly, and one half where you perform a super error-prone PCR (for instance, by spiking in mutagenic dPTP and 8-oxo-dGTP along with normal dNTPs; see Zaccolo et al) that gets you to 5-20% random mutations per base on each molecule. To call homopolymers, assemble the non-mutated sample reads as normal, then align the mutated reads to this ‘reference’ consensus. The random mutations in epPCRed reads break up the homopolymers with sequence variations that the basecaller can ‘see’ on each read, but that average out of the consensus. Haven’t seen anyone actually implement this yet, but no reason it shouldn’t work in principle!
Yes, computing and molecular biology are the most advancing areas during of my life. Only 2 of restriction enzyme found at my school hood. Thanks for great shows.
Very professional, congrats for this video. I would add the amazing story of the huge decrease of genome sequencing cost starting at 3 billions$ in 2003 to 1000$ today and probably less than 100$ in less than 5 years from now for the most advanced machines.
Great video as per usual. One thing that you might find interesting is that you need much more than 1x coverage of the genome to get the full genome even without concerns about accuracy. That is because you are sequencing the genome randomly, so that it is like sampling with replacement. Imagine you have a bag of 100 marbles and you grab a marble at random and toss it back in and repeat, you will need to repeat that a lot more than 100 times to have picked each marble at least once or even 90% of the marbles.
The adorably named product is actually written MinION and is supposed to be pronounced min-ion (like the charged particle). I was beta tester in our lab before they launched their product. That was, when the first movie came out. It did not take long, before the sequencing device itself had a little pill shaped yellow cartoon character sticker on its surface and the Computer which was picking up its data had a similar background on its desktop. I have been sequencing with illumina machines ever since, but on scientific congresses, I still tend to visit the Oxford nanopore booth, sometimes catching myself pronouncing the sequencer as 'minion'. It is still an interesting product. The type of sequencer you hand over to a medical doctor or biologist together with a box of reagents on dry ice and some equipments for DNA preparation. Then let him jump out of a plane in the middle of a nowhere so he shall do sequencing of a novel Ebola strain.
Many years ago, when doing my thesis, I was in contact with that technology. Congratulations on the video! One small addendum, though. Although of course not directly mentioned, NIH's use of "an ABI 370 sequencer for the Human Genome Project" could be misleading for somebody not familiar with the Human Genome Project. The Human Genome Project was a multi-national project, which involved the bigger part of all relevant institutions in the field, run for 5 years and cost around 150 Million USD. Please forgive inaccuracies, I am citing from top of my head and it is a long time ago. It was cutting edge applied science at its time and far beyond what one institute with one type of machine could have done. The second and more important aspect is the performance scaling. It went in roughly 20 years from cost of 150 million USD to 1000 USD and from 5 years to 48 hours!. That's outperforming Moores law, just only few people recognized it.
Very good. You forgot to mention multiplexing. It is easy to mix samples from multiple people together with a small tag to identify which patient the sequence came from. This allows us to take advantage of large single run sequence “depth” (gigabases/run) to provide breadth. I know it is typical to multiplex 100s of samples in a single sequencing run.
that was amazing video :) decent length too! Thanks Jon. Hey, i'm the same with depression. Just keep the thought-> "It won't last forever. It'll ease off" and it always does. 🤗
Great stuff man! I remember having a lecture about this technique during my bachelor. But this 23 min video was probably more informing and interesting than that lecture
Your videos are so different from others... I'm always very impressed by your insight into all those technologies. Can't imagine how many hours it took to make this one. Kudos!
What are the connections between fast sequencing and personal identification? DNA sampling for forensics is well established. Can/could these (quick, portable, accurate) machines make identification from ADN samples viable at points of entry?
As someone who works in the clinical trial industry, I have seen these machines first hand and they are impressive. What is even crazier is this is like one of 80+ machines that all do similar types of diagnostics just not with DNA as obtaining DNA is kinda the main limiting factor as it isn’t as logistically easy to obtain as say a whole blood sample would be so while dna is good for a final test, it’s often better to perform cheaper and easier first rounds tests using immunology based testing. Would love to see you make a video on immunologically testing machines in future but it’s a very dense topic that is on the fore front on biotech
It's really amazing to me how many of the sciences and engineering fields have to come together to make all this possible. People that think humanity is getting worse discount things like this. It truly shows human potential and it's mind boggling everything we have had to figure out to get this far.
Im very interessted abot the future of Oxford nanopore. At the last time I gave a Genome to sequencing they generated long reads with the oxford nanopore and used it as a template to assemble the data from the high throughput sequencing. Lest see where this all ends up
After the breathtaking 4 order of magnitude reduction in genome sequencing cost of the 2000s, the price per genome has stagnated at $1,000 for nearly a decade now. In order for the full benefits of individually taylored 'personal' medicine to be fully realized on a meaningful scale, the cost needs to drop another 10^4 so that $1 whole exomes can be done in your doctor's office in the time you're sitting in the waiting room before being seen.
Thanks for this video, fascinating stuff. I support these amazing instruments in a genetics lab where they are used for cancer research. The IT requirements of these instruments are causing most networks to burst at the seems because of the high data output from the scans you mentioned. Each scan can be about 1TB of data sent on local LAN switches to a processing server and to local NAS storage. The impact of offsite and cloud back up is a monster on the WAN.
BAFFLING. You are amazing!!! I love your videos, but i can not judge how accurate they are... i just assume you know what you're talking about. Hahaha. But with this one, i have certain affinity with the field and... omg, you nailed it! While watching, i even jumped to like... "omg! I can suggest a better development to this argument he presented! Id be happy to contribute!" Only to realize you _go over_ the point that i'd add on the next breath. Really cool how in-depth you can be. You went through some veeeery field specific stuff. Quite cool. =)
To pick your interest, there is a Mass Spectrometer (more than one, actually) that measures mass-to-charge ratio within parts per million accuracy. One of wich "ignores the velocity distribution" of ions. (Im intentionally throwing words in the hopes you're like "omg what? That must be very cool!") Ahahahah
The sequencer is not measuring DNA variation, but only sequences DNA fragments. All these fragments then need to be aligned computationally to the template human genome. Only then you know whether you have SNP (pronounced as snip).
They sell the hardware, the software, the reagent kits, and it really is a full body business. Illumina HQ is also literally across the street from UC San Diego and collaborates with UCSD bioinformatics on research, like every UCSD first authored bio paper I read in undergrad (ok more like skimmed thru tbh 😅) always cited Illumina Kinda cool fact, Illumina reagents are manufactured in Singapore!
Not just them, the ISeq, MiniSeq and NextSeq 550 are manufactured there aswell. We send them a bunch of new readheads every week from jenoptik, germany. ^^
Thanks for another great video. Exciting times ahead in the sequencing space, I'm particularly curious about the new 3D sequencing approaches which allow both higher throughput and new novel applications.
Thanks for a great whistle stop tour of sequencing! I've been reading up on sequencing as an alternative to STR alleles in forensic work. I'm not sure it is quite there yet. I was told that repeats represent a challenge for sequencers. So much so that people building a library of genomes have to do a separate sequence for repeats using older technology. I'm looking at Minion technology. It would be great if you could sequence the forensic alleles (e.g. the DNA-17 group) in a Lawyers office.
Thank you very much for this video! I very much hope you might do some more videos on the topic, maybe looking into the different technologies and/or devices presently used in even more detail :)
Illumina is the classical case, with their short read method beating the long read ones like 454 and PacBio, but these new nanopore companies do work a lot faster and in some cases cheaper. Perhaps after almost 50 years we will finally say goodbye to sanger sequencing methods
Does Illumina incorporate Crispr technology by chance? It seems like a logical application for a DNA sequencing company and can extend to genetic engineering
Crispr allows you to selectively edit a genome - Illumina machines allow you to read and sequence. One is writing the other is reading, and the CRISPR tech is patented by another group so while they could potentially incorporate it in some future machine they would have to pay for the license to do that.
@@jinngeechia9715 verification could be done with a standard PCR most of the time or even a simple sanger sequence, I wouldn't see the point of doing a full sequence unless the process becomes excessively cheaper. I might be missing something though
@@cianrynne2140 for a simple experiment, you don't need NGS to verify edits. For really serious application of GWE, you need it. These ambitious GWE projects will always have deep pockets for NGS.
Complete Genomics drops genome sequencing price to sub $100 at AGBT general meeting. DNBSEQ-T20×2, breaks the boundaries and reads the sequence of whole human genome for sub $100.
Appreciate the herculean effort of attempting such a summary. However, I'm afraid the video is not as comprehensible as most of your excellent other work. I think you should have rather gone step-by-step building several videos on simple concepts and technologies, and using more to the point graphics. For example, even the early history of Abbe and Zeiss microscopes in the 1850s that allowed the discovery of the cell nucleus and division is fascinating and worth telling,
Towards the end of the video I saw the Singular G4 and visit their website for some research - amazing piece of technology. Blown away how far we have come in benchtop sequencing over the past 20 years - Thank you Area 51 for the seed of that Alien Technology you found back in the 80(s)
I remember first hearing about illumina at a career fair as a fresh graduate back In 2013. I listened to their talk and was wowed by their tech. As a poor student I remember texting my dad immediately after saying he should buy their shares. Unfortunately he didn't.
I heard about the technology in the end of 2008 and instantly decided to build my career on analyzing the vast amount of data these machines pump out. Today we're talking about up to 2 TB raw data per sequencing run. Was kind of difficult at first but I found my way into the industry. I was a molecular biologist at that time and it was one of these moments where I just knew that this was going to take of big and very soon, so I wanted to jump on board. Short time after that, I got my hands on 454 and Solexa sequencing data, fitting right into my PhD thesis. Right spot at the right time and keen senses, that's the ticket ;-) Whom am I kidding.. It was sheer luck. I was slacking off on a hacker conference when some guy from the Sanger institute in the UK gave a talk about his work.
Masterful. I'd like to see the lowdown on modern medical analysis technology. How do they turn around a huge list of analyses of my fluids in less than a day?
Soon we will be able to play around with experimental DNA research in Labs. Like in Half Life. My imagination could picture a laboratory that studied different aspects of DNA research. Learning about specific gene, learning how certain parts of DNA could be edited. I've also heard that their is a form where they can try to use other nucleic acids in DNA to find stable combos of what they call XNA's.
Oh, turns out I was wrong! My apologies! Illumina bought Edico Genome's Dragen solution, which was a competitor to genalice. Genalice seems to have filed for bankruptcy in 2019.
Do you think you could do a video on the current state of Thorium reactor technology and infrastructure? I know China has finally gotten some operational and Germany has a great number of uranium plants they might convert.
The greatest value of Asianometry videos, for a long time now, was authors ability to present factual information in at least seemingly contextual way. Sadly this is yet another video of almost the opposite. Very interesting topic and a heap of disjointed potpourri technicalities that leave you with nothing in the end.
Sponsored by Blinkist. Get 25% off Blinkist premium and enjoy 2 memberships for the price of 1! Start your 7-day free trial by clicking on this link: www.blinkist.com/asianometry
Is there a Moore's Law for genome-sequencing?
What do you think of the size of economy, US vs China?
Whose economy is bigger?
I was surprised to find out that the US economy is probably less than half the size of Chinese economy.
The sales of luxury cars in 2021:
Benz: 758,863 sold in China, 276,102 in the US
BMW: 846,237 sold in China, 336,644 in the US
Audi: 701,289 sold in China, 196,038 in the US
Porsche: 95,671 sold in China, 70,025 in the US
EV: 2.8 million in China, 0.4 million in the US
I looked into further, and there are lot of other evidences that prove China is bigger, has lot more money and resources than the US does.
Thank you for this clearly delineated and distinct sponsor read, much nicer than the fractional art investment thing in a recent video.
@@sbkarajan What does this have to do with the video? Just cos it mentioned one of their competitors is Chinese?
@Asianometry Please stop commending Nobel prizes, it's a scam as most prizes and awards
Imagine if they merged with Texas instruments we would have the Illumina-Ti
"Illumina-TIE" NVidia, Jensen Huang
Well done sir
Boy are we a nerdy, 😂😂
and the absolute worst of the worst in filing ridiculously offensive lawsuits combined 🥴
Ha ha ha ha... 😎
One simple nitpick at the beginning: at the moment, next generation sequencing such as illumina is no longer called NGS since they’re … current generation. They’re called high throughput sequencing since 2018. Next gen sequencing are called long read sequencing, to avoid this exact problem
SNPs are often not spell out but pronounced as “snip”
As for 3rd gen sequencing, or long read sequencing, the field has pretty much settled on either pacbio or ONT. Hot topic right now is actually hybrid sequencing, where long read and short read are combined to improve both coverage and accuracy of read, as well as remove the bias introduced by the ref genome. We need to dive a bit deeper into how genome assembly works here, but in short we often align reads to a reference genome (from the human genome project for example). Now we can use the long read from Nanopore or PacBio as the reference, and assemble the short reads to this ref.
A point about the consumables for illumia: the reason we need way more nucleotides than present in the genome is because we underwent the amplification process. If we have an amplification factor of only 10 for all basepair, that bring number of nucleotides to the 10s of billions range. Account for waste and inefficiency, the consumables rate is reasonable.
God I hate SNPs as they "stole" the acronym from my company's stock ticker which confuses google
What about the sci-fi sequencing where you use nano-machines and SEM to actually read tiny bits of DNA (without PCR amplification)? How's that going? A decade or two ago that was the "next big thing". Haven't kept up with the field since then.
@@raylopez99 that sound like nanopore technology (ONT). It’s going strong! They have a version (MinION) that about the size of an external SSD. Super neat for field works and environmental science folk.
As a computational scientist, id say they have their problem with bias and lower accuracy than amplification methods, but they can be supplemented with short reads to build a robust and cheap sequencing
Everyone I work with simply refers to Illumina as NGS. Perhaps as consumers rather than developers of sequencing technology we’re too disconnected or lazy to bother making a distinction. Though it does speak to Illumina’s dominance in short reads. I agree that ‘next generation’ has been a comical misnomer for a decade at this point.
@@VuLamDang ONT's long read quality is better than you think. I routinely get quality scores comparable to illumina, every year they make huge strides on that, and now they have the advantage of being able to detect modified bases much better than any other method. If your experiments need long reads and nucleotide modifications, then ONT is really the only way to go.
Hi Jon! I am a sales specialist for Illumina Taiwan, delighted to see your video on Illumina! I have been following your channel for a while, really enjoyed the story about Taiwan’s camphor production history!
Hi Thomas, greetings from JenOptik ;-) We'll send a bunch of new MNSQ & NSQ readheads out of the manufacturing to singapore tomorrow, make shure you'll sell them well. ;-)
Hope you enjoy those exorbitant stock options while the company is in severe distress!! I hope the owners stop paying so those who are corrupt leave or get FIRED!!!
I appreciate how you described the process of sequencing and the history that led to instrumentation. I was trained on Illumina HiSeq 2000s, MiSeq, MiniSeq, and NovaSeq (briefly) when I was a molecular technologist. They're easy to use, and technically interesting on how they operate. The modular design of their more recent line of sequencers helped with cost and troubleshooting, and when I was a supervisor (I'm in Pharma testing now) this helped decrease the pricing for patients. Easily could design testing under $300 for most panels, and cheaper with higher volume.
Funnily, the HiSeq 2000 used readily available microscopes as the core body, they wrapped motorised parts around it. A lot of components are off-the-shelf systems integrated together.
Despite the model being phased out and recycled, many parts are useful for optical builds to this very day. I pulled a couple apart and almost all the components have been reused at our institution.
Recycling, done right!
As someone who uses these machines for work (don’t dox me bro), I’m amazed how well this was described. Great job.
Same here.. I work as an IT consultant and support these in a genetics lab. The IT requirements for these is a monster.
why u using your name and address as your avatar if u no wanna dox?
@@timmata8143 Its ok. Its just some additional money you have to spend. But I remember the reaction of one of our colleagues in the IT department, when I requested some computing resources. He proudly offered me a VM with 16(!) Threads and up to 16 GB RAM. So I said: You did nothing wrong, everything is OK, but I need to talk to your boss. ;-)
Right now we run two servers with a total of 1.75 TB RAM and 256 Threads.
Do not get me started on storage. Sore topic.
Wow. I cannot imagine how much time he puts in to do the research, find the photos, edit the video, recording, and organizing, to create this video. Good job.
Imagine you could use google:)
@@orangestapler8729 It really is master level. You can't show me a single teacher or professor researching and making a material like this for their studies let alone the delivery.
It's missing the Polymerase syntesis, that allows to indefinetly grow the gen sample (the "amplify" of the gen sample you talk about, is done with it), discovered by Margarita Salas, and that allowed the start of the human genome project 1 year after
Was Margarita Salas PCR discover ?... but He have a very good trips.
I am addicted to this channel now, kudos to you. Like the topics you choose and the way it's presented
@asianometry, I just want to say 'THANK YOU' for the work you do on your videos. I always learn a lot watching them, and really enjoy your straightforward, thorough coverage of whatever field you're looking at. And your presentation is great - not going off into tech-speak, but not talking down to the viewer either. I learned a lot about gene sequencing from this video, and have a bunch of things I now want to go look up to learn more. Well done!
I really love how well detailed you've explained these concepts, for a gu your background, people find it hard to grasp. Great content . Molecular biologist here 🙂
I think it's amazing how we don't know yet how profoundly DNA sequencing will effect our relationship with our own bodies and lives as a society.
I do a lot of DNA sequencing for my work, and really enjoyed your recap of the history of sequencing technology.
A fun potential future video topic in this area: will nanopore sequencing, and Oxford Nanopore specifically, supplant Illumina and optical sequencing-by-synthesis over the coming decade? A custom ASIC is at the heart of Oxford Nanopore’s flow cells, their CTO originally worked at Solexa, and their products seem (to me) to be right on the cusp of surpassing Illumina on key sequencing quality and quantity metrics.
At the time when I used them, sequencing of homopolymers was a bit of an issue with the Oxford Nanopore MinION. Similar to the iontorrent. Back then the algorithms would sometimes miscount the number of nucleotides in this type of sequences.
@@johnsmith1926 I'm curious as well how conventional sequencing avoids misinterpreting long stretches of one nucleotide. Wouldn't synthesis proceed right through that sequence and the fluorescence not be able to distinguish it's length?
@@samrusoff from my understanding, the florescent molecule prevents further nucleotides from binding to the RNA strand being synthesised, meaning that for any given cycle, only one nucleotide can join to the end of a strand. The remaining reagent is cleared away, image of the clusters taken, then a buffer is pumped through the flowcell, destroying the florescent molecule before another set of reagents is pumped through to begin the next cycle.
@@johnsmith1926 Homopolymer basecalling remains a challenge, but not an insurmountable one. ONT has switched to nanopores with longer channels, so they read a longer stretch of nucleotides and can accurately basecall up to 10-20 bp homopolymers. And in the last year they’ve tweaked their flow cells so that for a large fraction (up to >50%) of the DNA molecules they read, they thread both strands of the duplex through; and by basecalling with information from both strands of the duplex they can get to Q30/99.9% accurate single molecule reads.
This means that for applications where exact basecalling of long homopolymers matters, it should be possible to use a ‘polishing with dirt’ strategy, where you split and barcode a sample into two: one half that you prep and sequence regularly, and one half where you perform a super error-prone PCR (for instance, by spiking in mutagenic dPTP and 8-oxo-dGTP along with normal dNTPs; see Zaccolo et al) that gets you to 5-20% random mutations per base on each molecule. To call homopolymers, assemble the non-mutated sample reads as normal, then align the mutated reads to this ‘reference’ consensus. The random mutations in epPCRed reads break up the homopolymers with sequence variations that the basecaller can ‘see’ on each read, but that average out of the consensus. Haven’t seen anyone actually implement this yet, but no reason it shouldn’t work in principle!
What sources would you recommend to better understand ONT?
Love this. Quality checks of 'ngs DNA sequencing' are so standardised using illumina reads that it's used as a basis for everyone else's.
Keep doing whatever interests you.
WE WILL WATCH!!!
Yes, computing and molecular biology are the most advancing areas during of my life. Only 2 of restriction enzyme found at my school hood. Thanks for great shows.
Let me guess. One of them is EcoRI?
@@johnsmith1926 Yep, Escherichia coli is the most popular bacteria in biological study.
Lucifer asa
Very professional, congrats for this video. I would add the amazing story of the huge decrease of genome sequencing cost starting at 3 billions$ in 2003 to 1000$ today and probably less than 100$ in less than 5 years from now for the most advanced machines.
Nice digest of DNA sequencing and SNP analysis. Thanks!
Great video as per usual. One thing that you might find interesting is that you need much more than 1x coverage of the genome to get the full genome even without concerns about accuracy. That is because you are sequencing the genome randomly, so that it is like sampling with replacement. Imagine you have a bag of 100 marbles and you grab a marble at random and toss it back in and repeat, you will need to repeat that a lot more than 100 times to have picked each marble at least once or even 90% of the marbles.
The adorably named product is actually written MinION and is supposed to be pronounced min-ion (like the charged particle). I was beta tester in our lab before they launched their product. That was, when the first movie came out. It did not take long, before the sequencing device itself had a little pill shaped yellow cartoon character sticker on its surface and the Computer which was picking up its data had a similar background on its desktop. I have been sequencing with illumina machines ever since, but on scientific congresses, I still tend to visit the Oxford nanopore booth, sometimes catching myself pronouncing the sequencer as 'minion'. It is still an interesting product. The type of sequencer you hand over to a medical doctor or biologist together with a box of reagents on dry ice and some equipments for DNA preparation. Then let him jump out of a plane in the middle of a nowhere so he shall do sequencing of a novel Ebola strain.
This was great! I joined the Oligator team at alumina in 2001 right out of university fantastic company and leadership!
Many years ago, when doing my thesis, I was in contact with that technology.
Congratulations on the video!
One small addendum, though. Although of course not directly mentioned, NIH's use of "an ABI 370 sequencer for the Human Genome Project" could be misleading for somebody not familiar with the Human Genome Project. The Human Genome Project was a multi-national project, which involved the bigger part of all relevant institutions in the field, run for 5 years and cost around 150 Million USD. Please forgive inaccuracies, I am citing from top of my head and it is a long time ago. It was cutting edge applied science at its time and far beyond what one institute with one type of machine could have done.
The second and more important aspect is the performance scaling. It went in roughly 20 years from cost of 150 million USD to 1000 USD and from 5 years to 48 hours!. That's outperforming Moores law, just only few people recognized it.
"now on with the show" I missed that line. Thank you sir 🙂 fantastic channel.
I like the videos about genetics/micro biology. Good video as always.
Thank you and Much Love from the Philippines.
Very good. You forgot to mention multiplexing. It is easy to mix samples from multiple people together with a small tag to identify which patient the sequence came from. This allows us to take advantage of large single run sequence “depth” (gigabases/run) to provide breadth. I know it is typical to multiplex 100s of samples in a single sequencing run.
that was amazing video :) decent length too! Thanks Jon. Hey, i'm the same with depression. Just keep the thought-> "It won't last forever. It'll ease off" and it always does. 🤗
As someone working in molecular biotechnology, following this channel purely out of semiconductor curiosity, this was very cool to watch. Well done!
Thank you for all include ALL of the DNA discovery team.
Another gripping tale of invention and discovery. Bravo, sir! 😃
Great stuff man! I remember having a lecture about this technique during my bachelor. But this 23 min video was probably more informing and interesting than that lecture
Your videos are so different from others... I'm always very impressed by your insight into all those technologies. Can't imagine how many hours it took to make this one. Kudos!
Wow, your videos keep getting more and more informative. Thank you for all the work, and then sharing.
What are the connections between fast sequencing and personal identification?
DNA sampling for forensics is well established. Can/could these (quick, portable, accurate) machines make identification from ADN samples viable at points of entry?
Do a video explaining CRISPR technology
As someone who works in the clinical trial industry, I have seen these machines first hand and they are impressive. What is even crazier is this is like one of 80+ machines that all do similar types of diagnostics just not with DNA as obtaining DNA is kinda the main limiting factor as it isn’t as logistically easy to obtain as say a whole blood sample would be so while dna is good for a final test, it’s often better to perform cheaper and easier first rounds tests using immunology based testing. Would love to see you make a video on immunologically testing machines in future but it’s a very dense topic that is on the fore front on biotech
I feel stunned by the amount of people who barely understands one quarter of the content yet keeps watching every chapter. And enjoys them.
Illumina is by far the “Gold Standard “ when it comes to high throughput DNA sequencing
It's really amazing to me how many of the sciences and engineering fields have to come together to make all this possible. People that think humanity is getting worse discount things like this. It truly shows human potential and it's mind boggling everything we have had to figure out to get this far.
Im very interessted abot the future of Oxford nanopore. At the last time I gave a Genome to sequencing they generated long reads with the oxford nanopore and used it as a template to assemble the data from the high throughput sequencing. Lest see where this all ends up
After the breathtaking 4 order of magnitude reduction in genome sequencing cost of the 2000s, the price per genome has stagnated at $1,000 for nearly a decade now. In order for the full benefits of individually taylored 'personal' medicine to be fully realized on a meaningful scale, the cost needs to drop another 10^4 so that $1 whole exomes can be done in your doctor's office in the time you're sitting in the waiting room before being seen.
😄 that's never gonna happen fool. Be realistic. It can come back to 100$ though.
@@nishant54 there is nothing in physical law preventing it from going arbitrarily low, FOOL
@@Muonium1 Everything preventing it from becoming near free fool.
There's a service that offers whole genome sequencing for $400, in US -- I wonder what kind of equipment they're using
Thanks for this video, fascinating stuff. I support these amazing instruments in a genetics lab where they are used for cancer research. The IT requirements of these instruments are causing most networks to burst at the seems because of the high data output from the scans you mentioned. Each scan can be about 1TB of data sent on local LAN switches to a processing server and to local NAS storage. The impact of offsite and cloud back up is a monster on the WAN.
BAFFLING. You are amazing!!! I love your videos, but i can not judge how accurate they are... i just assume you know what you're talking about. Hahaha. But with this one, i have certain affinity with the field and... omg, you nailed it!
While watching, i even jumped to like... "omg! I can suggest a better development to this argument he presented! Id be happy to contribute!" Only to realize you _go over_ the point that i'd add on the next breath. Really cool how in-depth you can be. You went through some veeeery field specific stuff. Quite cool. =)
That said, i would LOVE to see your take on Mass Spectrometry and Omics sciences! (Mostly proteomics... hehehehe)
To pick your interest, there is a Mass Spectrometer (more than one, actually) that measures mass-to-charge ratio within parts per million accuracy. One of wich "ignores the velocity distribution" of ions. (Im intentionally throwing words in the hopes you're like "omg what? That must be very cool!")
Ahahahah
The sequencer is not measuring DNA variation, but only sequences DNA fragments. All these fragments then need to be aligned computationally to the template human genome. Only then you know whether you have SNP (pronounced as snip).
ASML is the new unit of measurement. Move over library of congress.
Love your videos. Would love to see more biotech and genetic engineering videos
Great explainer! Hope you can do more with bioanalytical instruments and techniques! Interesting stuff! 😎✌🏼
They sell the hardware, the software, the reagent kits, and it really is a full body business. Illumina HQ is also literally across the street from UC San Diego and collaborates with UCSD bioinformatics on research, like every UCSD first authored bio paper I read in undergrad (ok more like skimmed thru tbh 😅) always cited Illumina
Kinda cool fact, Illumina reagents are manufactured in Singapore!
Not just them, the ISeq, MiniSeq and NextSeq 550 are manufactured there aswell. We send them a bunch of new readheads every week from jenoptik, germany. ^^
Thanks for another great video.
Exciting times ahead in the sequencing space, I'm particularly curious about the new 3D sequencing approaches which allow both higher throughput and new novel applications.
Thanks for a great whistle stop tour of sequencing! I've been reading up on sequencing as an alternative to STR alleles in forensic work. I'm not sure it is quite there yet. I was told that repeats represent a challenge for sequencers. So much so that people building a library of genomes have to do a separate sequence for repeats using older technology. I'm looking at Minion technology.
It would be great if you could sequence the forensic alleles (e.g. the DNA-17 group) in a Lawyers office.
Great Presentation . Really enjoyed the topic . Cheers
The Primer movie poster was a nice touch.
Thank you very much for this video!
I very much hope you might do some more videos on the topic, maybe looking into the different technologies and/or devices presently used in even more detail :)
OOOOH that Primer (movie) reference! Love it! One of my favourite movies, there is a really good video about it here on youtube from LondonCityGirl!
Greetings from a certain Thuringian city with long optics history that played some role in the development of illumina’s top of the line products
The founder of America's first German settlement, Franz Daniel Pastorius, got his law degree in your fine city.
Well, at least 3 of them with some soon to come followups ;-)
I worked with Larry Bock, small world. On top of his bio startup work, he also created a science and engineering festival promoting STEM to kids.
The washing dishes part of the ad got me. That's exactly how I consume your videos 😂
I feel at least 10 iq points smarter after watching your videos.
This ad is a much better genre than Masterworks.
We really should be glad to live in an era of such scientific advancements.
Oooops. Picture of Allan Maxam at 11:28 is actually picture of William N. Lipscomb, Jr., Chemistry Nobel 1976.
Could you do a video about Oxford Nanopore Technologies and their MinION sequencer?
9:41 Blueprint, Factory, Catalyst, Bricks
Illumina is the classical case, with their short read method beating the long read ones like 454 and PacBio, but these new nanopore companies do work a lot faster and in some cases cheaper.
Perhaps after almost 50 years we will finally say goodbye to sanger sequencing methods
Does Illumina incorporate Crispr technology by chance? It seems like a logical application for a DNA sequencing company and can extend to genetic engineering
Crispr allows you to selectively edit a genome - Illumina machines allow you to read and sequence. One is writing the other is reading, and the CRISPR tech is patented by another group so while they could potentially incorporate it in some future machine they would have to pay for the license to do that.
Crispr is something you do to genetically alter an organism, while genome sequencing is about reading the genetic information from an organism
Gene writing and editing requires NGS to determine if the editing has taken place and also to see if there are any off-target editing.
@@jinngeechia9715 verification could be done with a standard PCR most of the time or even a simple sanger sequence, I wouldn't see the point of doing a full sequence unless the process becomes excessively cheaper. I might be missing something though
@@cianrynne2140 for a simple experiment, you don't need NGS to verify edits. For really serious application of GWE, you need it. These ambitious GWE projects will always have deep pockets for NGS.
Complete Genomics drops genome sequencing price to sub $100 at AGBT general meeting.
DNBSEQ-T20×2, breaks the boundaries and reads the sequence of whole human genome for sub $100.
Outstanding summary of Illumina´s meteoric rise. Kudos!
#Luciferine
I see this video and my first thought is of the company Oxford Nanopore
Appreciate the herculean effort of attempting such a summary. However, I'm afraid the video is not as comprehensible as most of your excellent other work. I think you should have rather gone step-by-step building several videos on simple concepts and technologies, and using more to the point graphics. For example, even the early history of Abbe and Zeiss microscopes in the 1850s that allowed the discovery of the cell nucleus and division is fascinating and worth telling,
it's scary the confluence between genomics and informatics and AI.
Towards the end of the video I saw the Singular G4 and visit their website for some research - amazing piece of technology. Blown away how far we have come in benchtop sequencing over the past 20 years - Thank you Area 51 for the seed of that Alien Technology you found back in the 80(s)
Great work as always, always wanted to know about illumina rise.
Lucifer Rise
I walked here not knowing anything on the matter. Walked out well lectured and curious. How do you do this?
I remember first hearing about illumina at a career fair as a fresh graduate back In 2013. I listened to their talk and was wowed by their tech. As a poor student I remember texting my dad immediately after saying he should buy their shares.
Unfortunately he didn't.
another instance of parents not listening to their kids. ah well.
I hope that you have told him what he turned down.
I heard about the technology in the end of 2008 and instantly decided to build my career on analyzing the vast amount of data these machines pump out. Today we're talking about up to 2 TB raw data per sequencing run. Was kind of difficult at first but I found my way into the industry. I was a molecular biologist at that time and it was one of these moments where I just knew that this was going to take of big and very soon, so I wanted to jump on board. Short time after that, I got my hands on 454 and Solexa sequencing data, fitting right into my PhD thesis. Right spot at the right time and keen senses, that's the ticket ;-)
Whom am I kidding.. It was sheer luck. I was slacking off on a hacker conference when some guy from the Sanger institute in the UK gave a talk about his work.
Glad he didn't listen. ILMN stock price has dropped to almost 10 years ago's level.
@@jasonhu7995 illumina share price was ~80 in 2013. Its 214 as we speak. What are you smoking?
A new video, a new thing to learn. :)
Masterful. I'd like to see the lowdown on modern medical analysis technology. How do they turn around a huge list of analyses of my fluids in less than a day?
Another case of a potentially world leading UK based company being sold before it's time with solexa. We are just US incubator at this point.
Excellent documentary!
My friend Jay works for Illumina.
Very important topic.
Wow, getting through all of this without mentioning the importance of PCR?
LOL Primer poster, love that film
Soon we will be able to play around with experimental DNA research in Labs. Like in Half Life. My imagination could picture a laboratory that studied different aspects of DNA research. Learning about specific gene, learning how certain parts of DNA could be edited. I've also heard that their is a form where they can try to use other nucleic acids in DNA to find stable combos of what they call XNA's.
Illumina also makes a really nifty FPGA platform for genetic data analysis.
Technically, they didn't make as much as bought the dutch company Genalice, who made the fpga solution.
@@cptbenjaminwillard ahh, thank you for the correction.
Oh, turns out I was wrong! My apologies!
Illumina bought Edico Genome's Dragen solution, which was a competitor to genalice.
Genalice seems to have filed for bankruptcy in 2019.
All straight forward till the legal complexities
speaking of no bake cheesecakes, i'm surprised it hasn't won a nobel prize yet
It’s located here in San Diego. So close to my house
Please make a video on the recent deal to stop the nl (asml) and japan from exporting latest euv machines to china
Do you think you could do a video on the current state of Thorium reactor technology and infrastructure?
I know China has finally gotten some operational and Germany has a great number of uranium plants they might convert.
This outsider says, “I’m surprised I never heard the words SHOT GUN”
The greatest value of Asianometry videos, for a long time now, was authors ability to present factual information in at least seemingly contextual way. Sadly this is yet another video of almost the opposite. Very interesting topic and a heap of disjointed potpourri technicalities that leave you with nothing in the end.
I wish I could send undergrad me this video.
Where does 10x genomics fit in this space ?
Video starts @2:30
Bacteriophages is an interesting topic... even bacteri has viral enemies
I still don't understand why Illumina doesn't acquire Pacific Bio (Pacbio).
Thank you for this video
2:27 begin
Fascinating