The beads that promised everything
Before the outrage, some context – because to appreciate why any of this matters, it helps to understand what we are actually trying to do.
For the past twelve years or so our lab has used a technique called replication profiling, or marker frequency analysis (MFA). The principle is elegant: rather than caring about the actual sequence of a bacterial genome, we use whole genome sequencing to measure how often different parts of the genome are copied during DNA replication. In growing cultures regions close to the origin of replication are present with more abundance than regions near the terminus, so sequencing read depth across the genome gives you a high-resolution picture of where and how replication is on average happening in a population. The actual sequence data is largely irrelevant. It is the read depth pattern that tells the story.
It is, admittedly, a slightly wasteful use of sequencing. But it is extraordinarily informative, and when it works well, the replication profiles are beautiful – smooth, clean curves that reveal the replication landscape of the entire genome at a glance.
For this to work, you need high-quality genomic DNA. And for the past twelve of so years we have used one reliable kit.
The kit that worked
Our workhorse was the GenElute Bacterial Genomic DNA kit from Sigma. It needed some optimisation when we first set it up, but once dialled in, it delivered. We always struggled with some chemical contamination that messed up our 260/230 readings, but the DNA was clean and our results were consistently beautiful replication profiles.
Then, in their wisdom, Sigma – now owned by Merck – decided to discontinue it.
How these kits actually work
For those unfamiliar with genomic DNA extraction kits, a brief explanation is useful here – not least because the type of kit matters for what follows.
Column-based kits, like the Sigma GenElute kit, work by exploiting a neat piece of chemistry. Bacterial cells are first lysed – broken open using a combination of detergents, enzymes and sometimes mechanical disruption – releasing their contents into solution. RNase A is typically added at this stage to degrade RNA, which you do not want in your final product.
The resulting soup is then treated with proteinase K to degrade most of the proteins (including RNase A) and then passed through a column containing a silica membrane. Under the right conditions – high concentrations of chaotropic salts such as guanidinium chloride, low pH and the presence of ethanol – DNA binds tightly to silica while most proteins and other contaminants pass straight through. Wash buffers then remove residual impurities, and finally a small volume of low-salt elution buffer releases the DNA from the membrane and into solution.
The silica chemistry here is worth noting: silica is simply silicon dioxide, the same material as glass. Older kits from Macherey-Nagel used glass milk – literally a suspension of fine glass particles – which works by exactly the same principle. The chemistry has not changed in decades; only the format has been refined. Columns replaced glass milk because they are tidier, more reproducible and the glass can interfere occasionally with downstream applications.
The one thing all column-based kits have in common: they need a centrifuge. Which every molecular biology lab already owns.
Enter the magnetic beads
When the GenElute kit was discontinued, I faced a decision. We had tried the Monarch kits – they work, no question, but I had never been fully satisfied with the yields from their plasmid extraction and PCR cleanup kits compared to alternatives. With our go-to kit gone, I decided this was the moment to consolidate and switch everything to Qiagen.
I have history with Qiagen. During my PhD, kits were expensive luxuries and we worked mostly with Macherey-Nagel kits. They worked often but not always. When I moved to the University of Nottingham for my postdoc, the lab used Qiagen kits exclusively, and the difference was immediately apparent. They worked far more reliably and consistently. Qiagen built a reputation on that reliability, and for good reason.
© Iren Grigoryan, 2026
So I contacted Qiagen, explained my problem of the discontinued GenElute kit and asked for their recommendation for a replacement. They suggested the MagAttract HMW DNA kit. HMW stands for high molecular weight, which matters for sequencing applications: longer DNA fragments give better results, and this kit is designed to preserve fragment length.
© Iren Grigoryan, 2026
The MagAttract kit uses magnetic beads rather than silica columns. The chemistry of DNA binding is identical – silica-coated beads, chaotropic conditions, ethanol – but instead of centrifuging liquid through a membrane, you add the beads to your sample, allow DNA to bind, and then place the tube on a magnetic rack. The magnet pulls the beads with the attached DNA firmly to the side of the tube. You pipette off the supernatant, add wash buffer, resuspend the beads, return to the rack, remove the wash. Repeat as needed. Then elute your DNA into a clean tube.
© Iren Grigoryan, 2026
It is an elegant system, particularly for high-throughput work. No membranes to clog, easy to automate, scales beautifully to 96-well plate formats. For sequencing facilities and core labs processing large numbers of samples, magnetic beads are often the format of choice for exactly these reasons.
The prices
The GenElute kit, which processed bacterial genomic DNA reliably for fifteen years: £155 for 70 preps.
The Qiagen MagAttract HMW DNA kit: £245 for 48 preps.
That is a meaningful difference. Some of it may reflect the genuine cost of manufacturing magnetic beads versus silica columns. Some of it is, let us be honest, the Qiagen premium. That is a conversation for another day.
But here is the thing about magnetic bead kits. You need a magnetic rack.
The rack
© RudolphLAB, 2026
The magnetic rack is, in principle, a simple piece of equipment. It is a block of aluminium or plastic housing neodymium magnets positioned precisely at the side of where the tubes sit, so that the magnetic field pulls beads firmly against the tube wall. The Qiagen version is well-engineered – it holds up to 12 tubes, transfers smoothly between the heat block and the rack, fits the tube format precisely, and does what it needs to do reliably.
Neodymium magnets are not expensive. Aluminium is not expensive. Machining a block to hold them in the right position is not expensive. The manufacturing cost of this item, if produced at any reasonable scale, is difficult to imagine exceeding a few pounds.
The Qiagen price for the magnetic rack: £718.89.
£718.89.
Yes, it lasts. Yes, it is well-made. Yes, it fits the system precisely. None of that changes what it is, which is a block of metal with some magnets in it, priced at nearly three quarters of a thousand pounds. It is, to use the technical term, a rip-off.
To be fair: you only buy it once. But you still have to buy it if you do not have the facilities to make your own.
Does the kit actually work?
Here is where honesty requires some nuance.
For routine genomic DNA extraction, the MagAttract kit works well. It produces clean, high molecular weight DNA in reasonable quantities. We have used it for PCR, for general sequencing applications, without complaint. It does what it says on the tin.
For whole genome sequencing specifically, Qiagen stand behind it – and rightly so. We have submitted samples extracted with this kit and received clean sequencing data back. In that respect, it delivers, exactly as Qiagen said it would.
The problem is our specific application: replication profiling.
We have submitted two independent batches of 24 samples extracted with the MagAttract kit for MFA analysis. In both cases, sequencing data came back without issue, but the replication profiles were not clean. Rather than the smooth curves we are accustomed to, we saw irregularities – and most tellingly, highly transcribed regions of the genome, such as the ribosomal RNA operons, were consistently under-represented.
We have seen this pattern before, with column-based kits, and we know what it usually means: protein-bound DNA is not being efficiently released during extraction. Regions with high levels of protein association – like heavily transcribed genes – are lost because the DNA does not bind properly to the extraction matrix. The fix, in our experience with columns, is to extend the proteinase K digestion step, giving the enzyme more time to strip proteins from the DNA before binding. When we did this with column kits, the under-represented regions disappeared entirely.
We tried the same approach with the MagAttract kit. No improvement.
As a control, I tracked down a remaining GenElute kit and submitted another 24 samples extracted with the old method. The results, however, were not straightforwardly clean either. Which rather complicates the narrative.
The honest conclusion is that I cannot cleanly blame the MagAttract kit, but I cannot exonerate it either. The slightly unusual features we are seeing in our sequencing runs may reflect something about the samples themselves, or the way the sequencing libraries were constructed, or factors we have not yet identified. What I am no longer confident about is that the kit is the primary culprit.
What I am confident about is this: failed or suboptimal sequencing runs are expensive. Very expensive. Once a kit has been associated with two problematic runs – fairly or not – the threshold for trying it again is extremely high. The cost of another failed batch is not something a lab of our size absorbs easily. For that reason alone, if I ever accumulate enough budget for further MFA experiments, I will probably reach for a column-based alternative first. Not because I am certain the MagAttract kit is the problem, but because I am not certain enough that it isn't.
Where we are now
The MagAttract kit has found a home in our lab for routine genomic DNA extractions. It works, it is convenient, and the format is well-suited to processing multiple samples. We will continue using it for those purposes, mostly because we have the outrageously expensive rack.
For replication profiling, the search continues. The GenElute kit is gone. The Monarch kit is a possibility we have not fully explored for this application. Other column-based alternatives exist and will be evaluated.
And the magnetic rack sits on the bench, doing its job, holding its tubes, pulling its beads to one side with quiet, expensive efficiency.
At £718.89, it had better last forever.
Similar and related blog posts: