Dr. Daniel Richards | UZurich

First Contact - 15.05.2025

Why did we contact them?

We contacted Dr. Daniel Richards, a Senior Scientist at ETH DeMello Group, to validate the viability of our proposed sequential design on a diagnostic paper strip, using distinct reaction zones, enabled by microfluidics, for sample preparation, RPA and CRISPR/Cas12. We also wanted to get feedback regarding DryLab being enthusiastic about using COMSOL Multiphysics to model the system.

Discussion

Microfluidics role in our project

Microfluidics is challenging for beginners. MSc Students under supervision of Dr. Richards always have difficulties with that, [First, after 2 years of studying microfluidics, they are able to design normal microfluidics chip/ paper] - he said. Moreover, a problem with using urine as a sample is the heterogeneity of the samples . However, he recommended us to focus on lab experiments at first, that would serve as proof of concept for us. For iGEM he would advise us to pay attention to visual representation of our idea, meaning we could use a 3D printed version of our paper-strip, looking the way we imagine it to be. It's important to show that model to others, people not from our field, and ask what they think about it: “does it look easy to use? Do you like the design? Would you use it? “ etc.

Detection method

Dr. Richards told us how problematic it is, to mix probes with RPA well enough. So, we should dilute it. He stated that we can use the RPA and the CRISPR-Cas in the same place. Most of the papers describe doing them one after another, but he did it with others at DeMello Group in one spot and still got good results. There was a very slight decrease in sensitivity & specificity, so that they still published the paper. RPA and CRISPR/Cas are widely used. They published a paper using this method for HPV. gRNAs can be found in papers, they also used the ones that they found in literature and they worked well.

If we want to add some novelty to our project, toehold-switch would be his recommendation. It was successfully used for detection of Zika virus and produced visible colorimetric readout as a result.Dr. Richards also mentioned that RPA alone can be used as a detection method, but did not specify how exactly.

Freeze-Drying Concerns

We want to Freeze-Dry reagents on our paper strip to make its shelf life long without a need for a cold chain. Dr. Richards approved that all RPA and CRISPR/Cas reagents can be freeze-dried. However, conditions should be chosen carefully to get the result that we aim for. He advised to flash freeze reagents with nitrogen and then slowly reheat it.

Modeling our System

Dr. Richards found our idea of using COMSOL Multiphysics too ambitious, since we don’t have any programming experience or background knowledge of using that platform. He explained how high the learning curve for this program is. We asked if he could recommend anyone able to guide us in the beginning or give us any kind of crash course, but he was sure that there was no one around him confident enough in using COMSOL Multiphysics to help us.

We then asked if Lucas-Washburn equation would be a reliable reference to use for modeling sample flow. He suggested that it would be a good idea to start with that.

Main Takeaways

Use 3D printer to present our idea visually
Make the design user friendly, by asking people outside our field to evaluate it.
RPA and CRISPR/Cas can run simultaneously with a minimal decrease in sensitivity.
RPA and CRISPR/Cas is a feasible method to detect STIs.
gRNAs can be found in literature and we can rely on them.
If we want some kind of novelty in our project, we could use toehold-switch.
All RPA and CRISPR/Cas reagents can be freeze-dried.
Use Python to model the system based on Lucas-Washburn equation instead of complex COMSOL Multiphysics.

Integration

We implemented Dr. Daniel Richards’ advice in the Dry Lab by using only Python to model our system. In the Wet Lab, his suggestion to add novelty guided the inclusion of an additional detection method—DNA hybridization—alongside CRISPR. One of our initial goals was to explore freeze-drying the prototype, which was a key reason for consulting Dr. Richards. However, as he noted during the interview, this process is highly complex, so we decided not to include it in the Wet Lab experiments due to time constraints and instead focus on developing a proof of concept. We also followed his recommendation to explore user perspectives by asking diverse questions about test preferences in a survey prepared by the Human Practices team.

Second Contact - 16.07.2025

Why did we contact them?

We contacted Dr. Daniel Richards, Senior Scientist from DeMello group at the ETH, for the second time in order to learn more about running RPA experiments. We wanted to prepare ourselves properly for the RPA experiments as no one in our lab has experience of working with such an amplification method. Moreover, we wanted to discuss methods to get a single-stranded RPA amplicon, because the DNA-hybridization detection method needs the RPA output to be not double-stranded.

Discussion

General things to look out for while performing RPA

The biggest challenge in RPA, according to Dr. Richards, is contamination. The reaction mixture is highly viscous, which makes pipetting prone to bubbles and errors. To minimize risks he advised us to:

Store RPA reagents separately from primers and targets.
Prepare everything else before opening RPA tubes.
Prepare negative controls separately, and handle them before positives.
Work in a clean hood, wipe surfaces with ethanol, and change gloves frequently.
Keep strict separation of pre- and post-amplification areas as post-amplification products pose the highest contamination risk.
Aliquot primers to avoid repeated freeze–thaw cycles.
Let solid reagents sit briefly before pipetting, and pipette slowly.
Expect variability as RPA often gives all-or-nothing signals rather than linear responses, especially at low copy numbers. Reliable curves usually require multiple experiments.
Include proper controls
- Negative controls are essential to rule out contamination.
- Positive controls ensure the kit and reaction components function as expected.

Moreover, primer design is critical. RPA primers are typically longer and GC content differs from PCR. In some cases, PCR primers may still work if melting temperatures are appropriate. Sensitivity can reach as low as ~10 copies, but performance often drops after lysis; ~100 copies is still good, while ~1,000 indicates poor sensitivity. However he argued that for diagnostic purposes such as STIs, qualitative results (yes/no detection) are more important than precise quantification.

Getting a single-stranded result with RPA

RPA normally generates double-stranded DNA, but for hybridization-based detection, a single-stranded template is required. Possible solutions discussed:

Asymmetrical RPA: Adjusting primer concentrations to favor single-stranded product. This is technically feasible, but sensitivity drops because the reaction spends longer in the linear phase rather than exponential amplification. It also requires redesign of primer melting temperatures, making optimization difficult.
Equilibrium approach: Instead of forcing single-stranded amplification, design the system so the probe is in large excess, driving binding to the target strand despite its double-stranded nature. This is often simpler and more reliable than asymmetrical RPA.

Testing RPA

qPCR is faster (~30 min) than running a gel and allows real-time differentiation of results, but can suffer from error rates; qRPA is a potential alternative. Using SYBR Green works well if amplification is clean, though primer dimers are a concern.
Molecular probes offer higher specificity but are expensive (~400 CHF per target).
Gel electrophoresis is cheaper but slower, only provides end-point data, and is less informative—especially for short templates (under 300 bp, requiring 2% agarose). RPA gels often appear messy, typically showing a target band plus primer dimers, so correct bands should be considered a successful outcome.

Lateral flow

Developing your own lateral flow strip is highly complex, requiring months of optimization across many parameters. Unless there is a specific reason, it is advisable to use existing solutions such as the Hybridetect universal lateral flow assay kit (from Milenia Biotec), where only primers need to be adapted.

Main Takeaways

Implement his advice on how to work in the lab with RPA to avoid contamination risk and get good results.
Possible methods to get a single-stranded template are either asymmetrical RPA or equilibrium approach.
2 methods to test RPA: qPCR and agarose gel electrophoresis.
Highly suggested buying the Hybridetect universal lateral flow assay kit.

Integration

Dr. Richards gave us a considerable amount of suggestions to avoid having problems with RPA, we implemented all of his advice to diminish the risk of contamination. That is working under a hood with gloves and being careful not to spill things, storing RPA kit in a different freezer than primers and DNA, running purification and gel preparation in a different bench than the hood used to run RPA, preparing aliquots and controls. We additionally researched more about asymmetric RPA and chose to follow it. As a testing method for the RPA, we decided to use agarose gel as no one in our lab has experience with it nor do we have the qPCR machine. The Hybridetect universal lateral flow assay kit, as suggested, was bought.