After Nanopore Sequencing Comes Quantum Pore Sequencing
I currently work for PacBio as a bioinformatist developing some methods to handle single molecule data and genome assembly properly.
Recently, I feel I am so lacking of vision. I have spent most of my time helping to develop methods in hope that they will be useful for the scientific community to use PacBio data. While we were developing those methods, as far as I could tell, many of those ONT fans had zero vision about them. We openly revealed those methods for the benefit to the scientific community to understand the value of PacBio’s and PacBio-like data. We naively assumed ONT would generate some great data with raw single molecule read accuracy > 96% as what Clive presented in 2012 AGBT. If so, those ONT fans would not need to use any of those methods we had developed. After a while, we find out that some of the visionary ONT fans are finally “inspired” to use some of our methods for processing some ONT data and publishing papers to show some values which some of those fan questioned about before. Without any new inspiration to me, I feel so dwarfed thinking how visionary the company was when they promised to build a world class bioinformatics center to develop bioinformatics method in Oxford back in 2011 ( https://www.genomeweb.com/informatics/oxford-nanopore- opens-new-bioinformatics-center-cambridge-will-expand-informatic).
Maybe it is just also from many other questionable scientific and business practices that trigger me to feel that I am in my mid-life crisis. I think one good way to get out of it is to learn to be a visionary man like @BioMickwatson. So, I start to think how I can start my own company to develop a quantum pore DNA sequencer soon.
Here is my vision: all sequencing and bioinformatics will be done with quantum computer with latest spintronics on a single graphane sheet. All you have to do is to put the graphane patch on the bottom of your Apple Watch, your genome sequence data will be transferred to the Cloud every 15 minutes (through HLTE, Hype-long-term evolution network, wireless throug hyperspace spectrum, of course). After analyzing with STP-WGAA (space-time-population whole-genome- assembly-association), the drug you need to keep your visionary view for the day will be on your office desk even before you get into your office in a Google self-driving car or through Elon Musk’s Hyperloop. Not only that, proper p-value for keeping your visonary view against whatever null hypothesis will be calculated and send to journal editors and reviewers to justify certain words used in the conclusion section of a news-worthy paper. Or, better, it also automagically produces papers because they are news-worthy as no scientific study is actullay needed to be done to get scientific results. (The best part is that it fixes my crappy English spelling and grammar in the process.) What a wonderful world!! The imagination is unlimited!! I will be retired in months after successfully hiring my CPO while many graduate students or postdocs will be figuring how to pipette into a graphene sheet or using a spintronics driving quantum computers in the years to come. (Did I say no pipetting nor bioinformatics needed?)
Life is too short. Now I feel inspired and visionary again. I should be hiring soon. First start with a CPO. What can be a more exciting job than a CPO!! Not surprisingly, my prototype already shows the p-value that this company will not be successful is less than 0.05. Join me! No resume of scientific research experience required. Experience for proper commercialization or production development undesirable. No one needs to write a product specification anyway. Rejection for sure if you try to make any scientific or business sense out of this. Also, if you already have public funding resource, our investors will prefer that you keep getting paid by that. High K-index impact guaranteed among cargo cult scientists’ club.
Opinion mine. Writing in a personal computer and in personal time. All Facts.
-——————————————————
Readers please rank the above vision in a 1-10 scale, with the following post measuring as 10.
Food, fantasies and the future
What will next generation livestock farms look like? Mick Watson examines scenarios and what we should do to get there.
Farmer Jane opened the gate and walked along the track that meandered along the side of her cattle barn. Chuckling to herself, she was old enough to remember how disease surveillance used to be done. It was so much easier now. Inside the barn, she approached the first of the ten cattle that had been randomly isolated, reached into her bag and took out the first of her SeqPensTM. Removing the protective lid, she briefly pressed the steel nib to the neck of the first animal then stood back to wait for the lights to change.
The painless microneedle pens would capture a tiny amount of blood and extract all of the DNA in the sample, whether from the cow, or the myriad parasites and pathogens that could infect her. The DNA would pass through an engineered nanopore that would determine the exact sequence in a matter of seconds.
-—————————————————
For those stuck in the mundane world of cost and quality, David Eccles wrote an informative reply regarding the questions asked by me and others. I am a bit confused about Reply 2 and need help from my knowledgeable readers.
1. Price is a bit steep at the moment, but that should change in the near future once ONT sorts out the issues with processing all the data generated in fast mode (see my update post for more information). Paying $900 for 400-2000Mbp of sequence is quite a lot, but the fast mode will produce that amount in about 3 hours (with a 48h run still possible). ONT is also planning to introduce a lower cost (around $300) for a 3-hour run time for the fast- mode runs. If cost is an issue, Id recommend waiting a couple of months for the output to increase.
2. Read error rate can be anything from 6% to 26%, depending on the alignment strategy and what you consider as errors. The 6% is using only 2D reads that pass the ONT read QC process (i.e. they end up in the pass folder), mapped with LAST using a custom alignment matrix, and only considering SNPs (its 11% including INDELs as well). The 26% is the worst mapping from ~20 runs using BWA-mem in ont mode, from all 2D reads. Both the 11% and 26% error rates came from the same set of data, 5 labs sequencing an E. coli K12 strain. Apparently PacBio INDELs can be fairly long; for my results the INDELs were typically 1-2bp, with 95% of INDELs being 8bp or less.
3. ONT said that explicitly that they dont want competitors to have access to their device through MAP. I think its also reasonable to assume that abusers of the technology will be prevented from further access as well. I havent asked anyone at ONT whether that would extend to people who are excessively negative about the technology Ill ask on the MAP wiki.
4. MinION metrics are really difficult to pin down. Read length doesnt make much sense; 8-10kb is a common target, but greater than 200bp would be my most precise statement, although thats probably going to change soon with an improved software workflow for short reads. Output is currently 400-2000Mbp for a full-length good run, but there are a lot of people in MAP who arent reporting on the community and are getting closer to 10-100Mbp per run (which I only found out from speaking to people at the conference). Note that the MinION can be stopped at any time, with the minimum reasonable run time to get sequence being about 15 minutes. All these metrics will change dramatically with fast mode (~30x more sequence), and again with the MkII chips (6x run time, 6x number of sequencing channels).
