16S Community Analysis - DIY or Outsource?

General
1

After we discussed how useful it would be to be able to identify different microbial species during last night’s journal club, I remembered that DNA sequencing is something that is possible to do at home.

There are two UK companies, [Bento Bioworks](https://bento.bio/) and [Oxford NanoPore Technologies](https://nanoporetech.com/), that make the equipment necessary to do mobile DNA sequencing, so that you can even do [DNA barcoding on a boat](https://bento.bio/story/barcoding-on-a-dive-boat-with-bento-lab-and-minion/):

  • - Bento Lab (£1299)

  • - [NanoPore MinION](https://store.nanoporetech.com/uk/minion.html) (£900)

    • [In short](https://bento.bio/protocol/dna-barcoding/), you use the Bento Lab's centrifuge to extract the DNA, it's PCR to amplify the DNA, and the gel electrophoresis to verify that it worked. Then you use the MinION to actually sequence the DNA.

    Obviously it's not quite as simple as that, but the point is that it's not out of reach for the home lab. There are open-source options out there for PCR(e.g. OpenPCR), centrifuges and gel electrophoresis as well. I'm not aware of an open-source nanopore sequencer though.

    2

    @“Gerrit”#p92 It’s taken me until now to work out who it was that I’d heard speak from Oxford NanoPore Technologies. We edited a video for the Water Action Platform way back during lockdown of a presentation that they made to Isle’s Technology Approval Group. I’ve now reached out to the presenter.

    Bento Lab looks really interesting too, I hadn't come across that before, but love their knowledge hub.

    Given our open nature, I'm always drawn to Open Science Hardware:

    ### Centrifuges

    [OpenFuge](https://www.instructables.com/OpenFuge/) with a $200 BoM back in 2013 initially looked promising but the inventor's last reply there was 10 years ago.

    The presumably closed source [SciSpin](https://sciquip.co.uk/scispin-mini-microfuge.html) Fixed speed 7,000 RPM / 2,680 g micro centrifuges are available for £90 and I might put centrifuges in the same category as autoclaves for open source sellers - too dangerous if something goes wrong.

    That said, Frank of Africa Open Science Hardware (with whom I had a great conversation after the last GOSH Roadmap co-working session) was [looking to develop one](https://forum.openhardware.science/t/call-for-collaboration-on-an-arduino-based-micro-centrifuge-for-rural-and-marginalized-areas-in-africa/4450/4). I'll ask him if he's made any progress or if the [FOSH Polyfuge](https://fosh-following-demand.github.io/Open-source-Centrifuge-for-WetLab/) (with [last commit](https://github.com/FOSH-following-demand/Open-source-Centrifuge-for-WetLab/) 3 years ago) is still the latest / best open source option.

    ### PCR

    [OpenPCR](https://github.com/jperfetto/OpenPCR)'s last software commit was 5 years ago, and the $499 buy now link on [OpenPCR.org](https://openpcr.org) now takes you to the $5,799 [Open qPCR](https://www.chaibio.com/openqpcr) replacement whose [last commit](https://github.com/chaibio/chaipcr) was 9 months ago.

    [PocketPCR](https://gaudi.ch/PocketPCR/) seems to be getting most love in the [GOSH forum](https://forum.openhardware.science/search?q=PCR) and has a €99 kit - I'm guessing from [their GitHub](https://github.com/GaudiLabs/PocketPCR) that it's more a product finalised by Gaudi 3 years ago than a thriving open source community.

    ### Gel Electrophoresis

    [Gel electrophoresis instructables](https://www.instructables.com/Gel-electrophoresis-system-mini/) similarly suffer from the last author response being 11 years ago.

    [openPFGE](https://gitlab.com/diegusleik/openpfge) looks more promising with the last commit only 1 year ago. They say "It costs USD ~$500" although their [2020 HardwareX article](https://pubmed.ncbi.nlm.nih.gov/35498240/)'s ~$850 may be more recent.

    Tobey [mentioned in the GOSH](https://forum.openhardware.science/t/pocketpcr-low-cost-usb-powered-open-source-pcr/2187/10?u=martinc) forum that IORodeo sold open source high-voltage supplies, gel-chambers and transilluminators - I'll ask Jo if they still do, or what they used to sell for.

    ### DNA Sequencing

    Is [ReSeq](https://forum.hackteria.org/c/reseq/19?ascending=false&order=activity) relevant?

    ---

    I think we should probably ask our friends at GOSH if I've missed the killer projects, like [OpenFlexure](https://openflexure.org) is for microscopy - with a massive active community & dedicated core developers.

    3

    @“Martin”#p102 I still have all the bits and pieces to build a centrifuge from the Biohack Academy. I just never got around to it as I didn’t have a use for it.

    Do you want to add your findings to the Amybo wiki?

    4

    @“Gerrit”#p108 yea, I should do - just not convinced we’ve reached any conclusions on any of the items yet…

    5

    I got a reply from Jo at IO Rodeo:

    >

    So, we stopped making and selling our gel electrophoresis and imaging line of products a few years ago. We ran into a sourcing issue with the blue LEDs we were using for the transilluminator. Basically, the quality control on the parts we received became unreliable which significantly increased the time we had to spend on manufacturing, quality control and testing. As we had trouble finding a good replacement part at a similar price point, we made the decision to stop selling them until we could work on a redesign. We also made the decision to discontinue making gel boxes and power supplies at the same time and focus on other areas of product development. We may bring back the transilluminator and electrophoresis power supply if we can find time to do the needed redesigns. We have all the links to the products on Instructables and in our archive which is currently not live. We are in the process of moving those to Github.

    6

    Trying to find the most cost effective way to determine what’s growing in our electroPioreactors. Please correct me if I’m wrong, but 16S community analysis appears to be the way to go.

    So far the best price I have found in the UK is £106+VAT per sample for:

    • DNA extraction
    • Library preparation
    • Sequencing
    • Data analysis and provision of raw sequencing data files

    OR, these are the best value prices I have found to do this in-house:

    Item Product Price (GBP)
    Centrifuge[1] SciSpin Micro Centrifuge £536
    PCR Machine[2] miniPCR Thermal Cycler $820
    DNA Extraction Kits[3] Qiagen DNeasy Blood & Tissue Kit (50 preps) £218
    Library Prep[4] Oxford Nanopore 16S Barcoding Kit (24 samples) £810
    Sequencer Oxford Nanopore MinION Starter Pack £1,799
    Data Analysis Qiime2/DADA2 £0
    Total ~£4,010

    That kit should provide enough for 24x 16S community analysis at £167/sample. Thereafter the consumables for 24 more samples would be 2x Flow Cells + 1x Library Prep + ½x DNA Extraction Kit = £2175 = £91/sample. So the breakeven with contracting it out (unless we can find a cheaper lab) would be 122 samples, assuming I haven’t missed anything critical and we don’t need to pay for training.

    What do you think? Can you find anything that’s better value for money? Should we start off with commercial analysis or try to find funding for our own kit?

    1. I have been struggling to find a safe opensource centrifuge, but just stumbled upon OpenCell when looking for open DNA extraction kits. ↩︎

    2. There are Open Source Thermocyclers like OpenPCR (was $499 until discontinued) or PocketPCR (€99 but limited throughput). ↩︎

    3. I’m not sure what reagents are required if we go for OpenCell ↩︎

    4. I beieve we could save on library prep by ~50% if using open protocols and buying reagents in bulk, but I’m nervous that I’d screw this up and waste the sequencing. ↩︎

    7

    If we screw things up, electrophoresis capability could be really useful:

    Item Product Price (GBP)
    Gel Electrophoresis Kit[1] Cleaver Scientific multiSUB Mini 10 System £262.00
    Agarose Powder Electrophoresis Agarose, A0169-10G £41.50
    Buffer Thermo Scientific™ TAE Buffer (Tris-acetate-EDTA) (50X), 1L £80.50
    DNA Stain[2] SYBR Safe DNA Gel Stain, 400 µL £65.65
    Transilluminator[3] Cleaver Scientific SAFEVIEW-MINI2 £397.00
    Total £846.65

    1. many open source options exist, I’d probably go with an easy Printables option or, while overkill for 16S, OpenPFGE looks tempting ↩︎

    2. we could use cheaper and/or non proprietary DNA stains ↩︎

    3. Many open source options exist like FluoPi and these on instructables and Hackster ↩︎

    8

    If you’re looking for a centrifuge, a PCR machine, gel electrophoresis and transilluminator, I think a Bento Lab as a four-in-one may be cheaper. They’re commonly used with Oxford Nanopores, see e.g. Rapid Water Testing with Portable Technology | Bento Lab (16s rRNA analysis)

    For data analysis, I’d suggest using the Galaxy Project, where you can run the analysis for free on their servers - it’s pretty amazing: Hands-on: 16S Microbial analysis with Nanopore data / 16S Microbial analysis with Nanopore data / Microbiome

    9

    Yea, GPT 4o seemed to have it in for the Bento Lab. It rulled it out based on centrifuge speed, but said the 7000g OpenCell would suffice. I’ve now seen it’s not comparing the Bento Lab Pro, however when it does, the savings aren’t as big as you’d imagine:

    Cost Comparison Table

    Component Original Custom Setup Bento Lab Pro Setup Maximally Open Source Setup
    Centrifuge SciSpin Micro Centrifuge £536 Included in Bento Lab Pro £1,995 OpenCell Centrifuge £50 insufficient 3000g for DNA pelleting Same as Original Kit £536
    PCR Machine miniPCR Thermal Cycler $820 Included in Bento Lab Pro openPCR ~£500
    DNA Extraction Kit Qiagen DNeasy Blood & Tissue Kit (50 preps) £218 Same as Original Kit £218 OpenCell DNA Extraction ~£50 + reagents ~£80
    Library Prep Oxford Nanopore 16S Kit (24) £810 Same as Original Kit £810 ~~£500
    Sequencer Oxford Nanopore MinION Starter Pack £1,799 Same as Original Kit £1,799 Same as Original Kit £1,799
    Electrophoresis System Cleaver Gel Box £262 Included in Bento Lab Pro DIY Open-Source Gel Box £50
    Agarose Powder Electrophoresis Agarose £41.50 Included in Bento Lab Pro Electrophoresis Agarose £41.50
    TAE Buffer TAE Buffer £80.50 Included in Bento Lab Pro DIY TAE Buffer Recipe ~£10
    Transilluminator Blue Light Transilluminator £397 Included in Bento Lab Pro DIY Blue Light Transilluminator £50
    Total Cost ~£4,790 (~£4,010 excluding electrophoresis) £4,822 ~£3,600

    4o now says “The Bento Lab Pro is compact but costs nearly as much as the Original Custom Setup without the same performance.” - I think what it misses[1] is the community/support elements of Bento Lab. What kills it though is if we decide that we should go as Open Source as possible, or that we can probably get the job done without electrophoresis.

    Galaxy Project looks amazing - yes whatever we do we should be able to run or rerun the bioinformatics there.

    1. as well as the $/£ exchange rate which makes the custom setup even cheaper ↩︎

    10

    Chai.bio, maker of the OpenPCR, have gone closed source. If we do want to go the full open source route, this would be another option: PocketPCR – The thermocycler for the rest of us

    11

    I just combined our topics on DNA sequencing so that everything is in one thread.

    I discovered this week that Plasmidsaurus is now offering 16S sequencing!

    For folks in the USA this is amazing news, as they offer extraction from raw samples. You can get a sample extracted ($15) and sequenced ($45) for $60 with a turnaround time of just three days!

    There’s also good news for us in the UK. While their London lab does not offer extraction from raw samples yet, you can send in purified 16S amplicons for sequencing for only $30, with a turnaround time of 1 business day. :exploding_head:

    Results include (full-length, because they use Oxford Nanopore) raw read sequences as well as taxonomy tables and plots.

    So @Martin , we need to figure out how to amplify and purify our samples! It seems like the Zymo Quick-DNA Fecal/Soil Microbe Microprep DNA extraction kit may be the way to go, as it is designed for low-throughput, portable workflows. For example, it only requires a 10,000g centrifuge (but OpenCell can only do 3,000g, so if we go down that route we’ll have to modify the protocol).

    I really like the Bento Lab Pro, but as pointed out previously it’s maybe a bit on the expensive side. I’m thinking a PocketPCR and OpenCell may be the cheaper solution here?

    12

    Great find @gerrit - crazy prices.

    Completely rewritten post, as the last version was rubbish

    I can’t figure out whether Plasmidsaurus want us to submit a purified PCR product or normalised genomic DNA - Gemini is suggesting £840-£1,870 capex plus ~£10.60 per sample for the former or £2,790 - £3,070 capex plus ~£23.10 per sample for the latter. I’ll email Plasmidsaurus for confirmation.

    If Gemini’s assumptions are correct and we can submit a purified PCR product we’d be looking at 22-50 samples to break even with outsourcing if it was available at the equivalent of $60/sample in the UK. Or if we must submit normalised genomic DNA 112-123 samples. The latter is pretty comparable to our 112 estimate from late last year, but if we can’t get a commercial service offering $60/sample in the UK, then DIY becomes much more attractive.

    I do however think that UK prices are likely to follow those in the states in due course, so I still think a public lab would be the best location for the kit (particularly if a fluorometer for normalised DNA is required), so that - once commercial prices drop, there’s still value in people being able to use the kit to learn how the sausage is made.

    13

    As soon as I saw that 16s sequencing was available I asked Plasmidsaurus when we’ll be able to send in raw samples in the UK - unfortunately they don’t have an ETA for that yet.

    That’s why I was wondering if there’s a cheap DIY method for doing this, so $50 OpenCell for DNA extraction, €99 PocketPCR for amplification, and either the $40 DIYNAFLUOR fluorometer or the Open Colorimeter Plus for DNA quantification. Of course that doesn’t include the cost of the reagents.

    14

    I’m amazed that a $40 open source fluorimeter can replace the standard commercial option at >£2,500, but it’s definitely worth trying. Perhaps step 1 would be to build one (or one of each) and check their accuracy and repeatability with something like this £88 dsDNA Assay Kit.

    If all is well then we invest in the rest of the capital kit, piece by piece:

    Gemini 2.5 Pro was suggesting we could get away with the OpenCell’s 3,000 g if we spun it for longer, but I don’t think that will work for the filtration steps. I’d like to think the 15,100g centrifuge I mentioned previously is <£550 could cover the 16,000g for 3 minutes required for ZymoBIOMICS DNA Miniprep Kit if we gave it an extra 30s. Or we could get something like a second hand Eppendorf 5415D and rent/borrow/build a tachometer to check it’s still hitting 13,200 RPM.

    We could of course find a friendly lab with a 16,000g + centrifuge and run a head to head with the OpenCell using extended times. By this stage our fluorimeter should be able to compare the DNA yields. If the OpenCell doesn’t suffice we can still hopefully use it in place of the recommended Vortex Genie. I’d probably want to check with Zymo before buying a sub 16,000 g commercial centrifuge and I don’t think I’d be happy going to that force with anything I made myself.

    Assuming everyone attempting this already has sufficient micropipettes, I think that’s the minimum kit covered. But to avoid wasted analysis costs (and also for our centrifugation head to head) we should check DNA length with Jo’s gel electrophoresis or Gaudi’s system, etc.

    I might be tempted to go with Plasmidsaurus’ Microbiome 16S Amplification & Sequencing starting at $45/sample rather than doing the PCR ourselves. If our labour was free we’d be saving money after the 8th sample, but given that I’d likely do a worse job than the labs, I’d rather just pay the extra ~£11 and save what I assume would be >1h per sample.

    15

    DIYNAFLOUR has a low-cost protocol for DNA assays at $0.13 per measurement.

    That’s exactly what I meant with:

    From the OpenCell paper:

    our testing demonstrates that only 2500 × g is sufficient for DNA extraction with some protocol adjustments […] For the centrifugation module, an optimal factor of 1.5x the protocol recommended duration is determined for efficient DNA isolation from three independent trials

    I suggested the Zymo Quick-DNA Fecal/Soil Microbe Microprep kit above because it only requires 10,000g - is there a specific reason why we would need the ZymoBIOMICS DNA Miniprep kit with a 16,000g requirement?

    I’m confused. I’ve mentioned previously:

    and

    and

    That was the whole point of my previous posts :joy: If the Microbiome 16S Amplification & Sequencing service was available outside of the US it would be a no-brainer, but I’m trying to figure out if there’s a low-cost way to prepare the samples to still take advantage of their London lab’s low-cost 16S sequencing and quick turnaround service.

    16

    On the $0.13 per measurement DIYNAFLOUR question. No, I hadn’t seen that. Seems great although I’m getting estimates of £230 to make the first batch - can you see cheaper routes? Its serial dilutions also introduce more scope for error than the £88 kit, so unless we can get the first batch price down, I’d be tempted to start with an AffiGEN or similar kit.

    My point of disagreement with the openCell paper was around filtration, while it made sense to me that an accelerated sedimentation step could be replicated with lower g and higher time, that isn’t alway the case with filtration. The final step in both Zymo protocols is “Transfer the eluted DNA to a prepared Zymo-Spin™ II-µHRC Filter in a clean 1.5 ml microcentrifuge tube and centrifuge at exactly 16,000 x g for 3 minutes.” (step 12 of the Fecal/Soil protocol) that is much harder IMHO than the final 16,000 x g dry spin step of the DNeasy Plant Pro Kits protocol that openCell adapted.

    TBH, the reason for my swapping out your suggested Zymo Fecal/Soil kit was more because we’re not using Fecal/Soil samples by this stage - even if they were the inoculum at the start of the electroPioreactor enrichment. Knowing the issue was the final 16,000 x g spin, I didn’t imagine they would be any easier than the regular kit - I thought, if anything, they would require more front-end work.

    Their Quick-DNA Fungal/Bacterial Miniprep Kit (50 Preps - D6005) however only require 10,000 x g - so in theory should be even easier for us to bodge an openCell protocol with than DNeasy. It won’t give as pure results as either the Fecal/Soil kit or the flagship D4300 kit which both include that final II-µHRC, but that’s the step I don’t think will work with the openCell. At (£247/50) it is also therefore cheaper than the Fecal/Soil kit which is cheaper than the flagship D4300 kit.

    I emailed Plasmidsaurus two days ago, as it unclear as to whether it’s just the “US-based labs can add extraction directly from raw biological samples for just an extra $15.” that we can’t get. That was what I assumed after a more detailed read, prior to that I assumed that the Microbiome 16S Amplification & Sequencing service was also US only. I was waiting for an answer before responding, but it looks like it won’t be coming this week.

    We could of course go with another lab. There are many commercial and university labs capable of this. What I think is special about Plasmidsaurus is that they publish unit prices, and are happy to offer those on single samples. If getting a quote from a UK lab, we’d probably want to batch up samples so we could send a feasible number for a nanopore flowcell.

    If we’re going with Plasmidsaurus though, Step 1 is the same whether we do our own PCR or not: build and test a fluorimeter.

    17

    They’re using SYBR Safe, which you’ll also need if you want to do gel electrophoresis, and LTE or TE buffer, which you need for storing DNA. That said, I agree that it’s more work than just buying a kit.

    I imagined we’d also like to sequence the raw samples, to see how the enrichment changes the microbial community.

    Ah, I missed that step, good catch!

    I’d add that their turnaround time is very fast, and they appear to be much cheaper. They also include taxonomy plots and tables with their microbiome sequencing results!

    18

    Wouldn’t you need to do serial dilutions with the £88 kit as well? Looking at the components it comes with 0 ng/µL dsDNA Standard #1 and a 10 ng/µL dsDNA Standard #2.

    19

    Gemini 2.5 Pro:

    That’s a very sharp observation, and it gets to the heart of why the kit is so convenient.

    The answer is no, you would not need to do serial dilutions with the AffiGEN kit.

    The reason is that the DIYNAFLUOR software and its accompanying protocol are specifically designed to perform a simple and rapid two-point calibration.

    Here’s how it works:

    1. The software first measures Standard #1 (0 ng/µL) to set the baseline fluorescence (the “zero” point).
    2. It then measures Standard #2 (10 ng/µL) to set the maximum fluorescence (the “high” point).
    3. It then draws a straight line between these two points. Any unknown sample you measure will have its fluorescence level plotted on this line to instantly determine its concentration.

    The protocol you provided for the DIYNAFLUOR confirms this method. It instructs the user to prepare only two standards for measurement, after which the software plots a “2-point calibration curve”.

    The AffiGEN kit is designed for exactly this workflow. It conveniently provides you with the two precise, ready-to-use points you need to perform the calibration without any extra work. This is a major advantage over the “from-scratch” protocol, where you would first have to perform a precise serial dilution yourself just to create the 10 ng/µL standard from a more concentrated stock.

    20

    Maybe it’s more helpful if you have look at the DIYNAFLUOR build instructions, page 35, where they use the official Qubit dsDNA kit from Thermo Fisher for validation.