Electrical detection of RNA cancer biomarkers at the single-molecule level | Hacker News Reader

Electrical detection of RNA cancer biomarkers at the single-molecule level

medicalxpress.com

107 pointsjdmark16 comments

Apologies for my enthusiastic interruption. I am deeply intrigued by this ctDNA work using electrical signal detection. However, I see two challenges with current early detection liquid biopsy methods. Firstly, they target signals that are inherently rare in blood (less than 1% of any blood volume). Secondly, they focus on either cfDNA or ctDNA, indicating that cancer has already progressed significantly or spread to other areas.

In this context, I strongly believe that our approach positions us as one of the most captivating players in this field. We aim to leverage the immune response as an indicator of the body's cancer status. By harnessing the highly effective detectors within the immune system, we actively pursue cancer detection. Additionally, relying solely on specific biomarkers seems reminiscent of manually curating features, akin to the early days of NLP when compared to modern transformers.

We are eager to share some of our methodologies and eagerly seek collaboration with professionals from diverse backgrounds, including genomics, wet lab scientists, bioinformatics experts, as well as AI and full stack engineers. Although we are currently in stealth mode and cannot divulge extensive details, we would be thrilled to engage with inquisitive and enthusiastic scientists. In some ways, we believe that our approach may also enable us to validate the efforts of the prominent players in the field of multi-cancer early detection.

For further discussions, please feel free to reach us at qubind@qubind.com.

> Firstly, they target signals that are inherently rare in blood (less than 1% of any blood volume).

That's one of the advantages of this technique.

> Secondly, they focus on either cfDNA or ctDNA, indicating that cancer has already progressed significantly or spread to other areas.

This isn't really accurate and is a bit misleading. For starters cf/ctDNA can be detected in stage I disease (and some limited reports of in situ). It's hard to compare studies or assess accuracy in early stage diseases as methodology is very variable and sample sizes aren't great yet. Ongoing trials will give more info.

The method described here isn't only for MCED testing, the more significant advancement is in disease monitoring and treatment response with the idea that we would be able to detect treatment failures and mutations earlier and easier with high fidelity.

Speaking about "liquid biopsies" and MCED in general, there still isn't really anything compelling to support this approach (for cancers that have screening tests available), we don't have nearly enough data to have an informed opinion on yield, benefit, and cost. Preliminary results have been underwhelming in terms of positivity rates, PPV and specificity.

> We aim to leverage the immune response as an indicator of the body's cancer status. By harnessing the highly effective detectors within the immune system, we actively pursue cancer detection. Additionally, relying solely on specific biomarkers seems reminiscent of manually curating features, akin to the early days of NLP when compared to modern transformers.

Does this actually mean anything? It reads as a bunch of marketing hype. Do you have a reference about the "highly effective detectors within the immune system" one might harness?

I think you’re spot on about the clear advantage of your approach relative to the article and blood biopsies: If you’re looking for cancer in the blood, you’re probably too late as this suggests metastatic spread.

You suggest something better: look for tissue level markers.

The immunotherapy revolution (ie CTLA/PD1 molecules), you’ll recognize that leveraging the immune system is brilliantly effective.

Presumably stealth genes do not evolve first, but rather proliferative genes. And this is why any chronic damage triggers cancer because renewal is a replicative process, and replications begets copy errors. wanderers and passersby will have to recognize I’m simplifying.

I have a great lateral. Not sure if you have the funds to try two things, but TCR characterization is a wonderful idea. At least thats the best way to do it, as I see it. has many application beyond cancer

Good luck!

> If you’re looking for cancer in the blood, you’re probably too late as this suggests metastatic spread.

Isn’t the point of cfDNA that you can pick up “dead” DNA fragments floating around long before you have viable cancer cells drifting through the bloodstream and lymphatic system?

Yes. Normal cells break apart all the time. If you measure total cfDNA amount in a blood sample, you'll find significantly a higher cfDNA concentration after you exercise. It's a normal part of the wear and tear of cells.

So, if you find cfDNA from a tumor, it's not necessarily indicative of metastatic spread... just that the tumor cells have been broken up. It could be from too much growth, or immune attack, or metastasis, or... etc.

cfDNA is great in that you can run a single assay to cover the entire body. cfDNA is difficult because your single assay covers the entire body. You can't localize the source of cfDNA.

So, they seem to have made special DNA probes that attaches to some cancer-specific bit of RNA, and the combination of the two stuck together is something they can detect via "scanning tunneling microscopy break junction (STM-BJ)". Is that right?

I guess the question is whether you can use this DNA/RNA detection method without having a whole sophisticated STM setup as the other part of the detector. The article says calls it a nanobiosensor, and says it "has a promising future as an inexpensive, highly sensitive and label-free miniaturized device for early-stage cancer screening of liquid biopsies".

This is really cool, but it's only a nanobiosensor insomuch as you can get a useful signal out of it without an expensive scanning tunneling microscope setup attached, I would assume!

I think you’re overestimating the complexity of an stm setup. There exist small coffee machine sized stm machines because the principal behind the operation is relatively straightforward. People have made diy stm machines using piezo electric buzzers as the component that moves a tungsten needle over a sample. It’s relatively easy to make atomically thick tungsten needle points because of how it fractures along its grain structure. The tech behind the principle of operation of an stm is already ubiquitous.

Basically. It's been a while since I've worked in wetlabs (clinical now) but looking at their methodology the fixed equipment costs would seem comparable to ddPCR and I'd expect cheaper than NGS currently used for ctDNA. Per-unit consumable costs also seem negligible.

The specific AFM setup they used here is pretty old and can be replicated (relatively) inexpensively.

I would guess the biggest cost barrier/disadvantage would be throughput rather than setup costs but these results appear to be a potentially good and/or cost-effective solution to the sensitivity issues.

Prefer: https://www.nature.com/articles/s41598-023-39450-6

Abstract:

>> Cancer is a significant healthcare issue, and early screening methods based on biomarker analysis in liquid biopsies are promising avenues to reduce mortality rates. Electrical detection of nucleic acids at the single molecule level could enable these applications. We examine the electrical detection of RNA cancer biomarkers (KRAS mutants G12C and G12V) as a single-molecule proof-of-concept electrical biosensor for cancer screening applications. We show that the electrical conductance is highly sensitive to the sequence, allowing discrimination of the mutants from a wild-type KRAS sequence differing in just one base. In addition to this high specificity, our results also show that these biosensors are sensitive down to an individual molecule with a high signal-to-noise ratio. These results pave the way for future miniaturized single-molecule electrical biosensors that could be groundbreaking for cancer screening and other applications.

Would be interesting if this could be adapted for long-covid too

They also made a electrical response test for me-cfs (which would include some kinds of long-covid) based on how cells respond to salt

https://www.nih.gov/news-events/nih-research-matters/blood-t...

This is neat, but seems wildly more complex than other detection methods with liquid biopsy. Most major players in this space are using methylation analysis with high specificity and sensitivity - it's not obvious to me that there is a benefit to electrical detection over that.

Not sure I follow, methylation analysis is very different from NA detection.

Why is this getting so many upvotes? Pretty unspectacular stuff.

Do you have a link to some similar method published before? For me, it looks interesting. But this is not my area, so I'm not sure if it's groundbreaking or just usual stuff with some irrelevant spin.

Well, what struck me was the upvotes to comment ratio, which hovered at n:0 for the longest time.

There is a similar posting just now (20:21 PST), also from medicalexpress.com. At this moment 71 points and 3 comments, https://news.ycombinator.com/item?id=37203761

Seems fishy.

My guess is that people love to upvote cancer cure, perhaps too much.

IIUC this is only a proof of concept, not a clinical trial, not even rats, so it's too early to know if it will be useful. I think it will be very specific because it can target an specific RNA, but I don't know if it can be adapted to test multiple markers instead of just one. Also, I don't buy the single molecule story. My guess is that have many (10^20) molecules in the solution but only one of them binds to the probe.

For an oficial reply, you can email the mods, so they can take a look.

Sometimes technical stories have a high upvote to comment ratio. It used to be common but now it's more unusual.

Reply on news.ycombinator.com