Perspectives

The experiments we performed were exploratory in order to obtain a proof of concept. We would need to significantly increase the number of replicates for the various results we obtained in order to have a thorough demonstration of our technology. Moreover, increasing the number of replicates will also strengthen our model fitting results (see below). In addition, with the idea of turning Shell’lock to a commercial product, we would need to enter more optimization cycles in order to fine tune the various aspects of the reaction. Indeed, as shown in the result section we quantified the sensitivity of our test. We showed that this sensitivity was above 10nM, which is a higher concentration than previously reported(1). To reduce this difference, we would need additional cycles of optimization and further development of the experimental procedure.

As mentioned above, we intended Shell’lock to be used by the primary stakeholders of the oyster production: the farmers. We showed in vitro that through lateral-flow assay the test could be easy to perform and could detect pathogenic sequences of microorganisms. However, we worked with synthetically synthesized pathogenic sequences. In the case of field detection we would need to implement an extraction step as previously described (1). Several strategies could be investigated and implemented in Shell’lock. Moreover, following the previous section, a key aspect of the test is its sensitivity. As our test targets RNAs we may encounter low concentration of the target. To address this problem we propose two different approaches:

• Increasing sample concentration to maximize the detection. A few milliliters of water sample could be passed through a 0.2µm filter so that microorganisms are concentrated at the filter surface. RNAs could then be extracted directly from the filter by adding a small volume of hot lysis buffer. This will increase the concentration of the available RNA.
• Amplification of the target before detection. This step would require more experimental characterization. As our test would be in the field, this reaction would need to be performed at room temperature, which can be achieved by techniques such as RPA (Recombinase polymerase amplification)(2). This step combined with T7 transcription would increase the concentration of the target, thus raising the sensitivity threshold.

Finally, as we target RNA in the sample we would be confronted with accessibility of the target sequence. Indeed, RNA can adopt 3D structures making base-pairing a difficult process. However, the synthetic sequences that we have produced and tested are around 300 bases long. As they can activate Cas13a, it is likely that there are no inhibiting secondary structures in the vicinity of the 28nt sequence.

As stated in our modeling page, our perspective would be to access mechanistic information about the Cas13a reaction. Indeed, we were surprised to observe that it can handle a couple of mismatches between the target and guide RNAs. With our model we can dissect how mutations affect each binding affinity and catalytic activities. Further experiments would include more mutations, varying their location and the type of mutation in order to have more insights into the mechanism of Cas13a activity.

The goal of Shell’Lock is to survey the appearance of pathogens ahead of infectious episodes. However, one key aspect of epidemic spread is containment strategies or solutions when the infection started. An additional objective would be to propose containment solutions to the oyster farmers. For this we are thinking of developing ecological/epidemiological modeling based on the typical compartmental SIR model(Susceptible, Infectious, Recovered). We would like to build a spatial ecological model that, informed by time series of on-field data obtained with Shell’lock at different locations, could provide containment strategies of isolation of small areas of the farm, without the need to lockdown the entire production line. The model could also inform about the best timeline and frequency for testing.

We showed in our result section that Shell’lock could be used to detect V.aestearianus sequences. But as shown in figure 1, we also designed targets against other types of organisms. Namely, we detected: oyster immune system genes, mussel genes and trigopus genes. These targets were not randomly selected and come from the idea that as an additional information given by the test the oyster farmer could probe the environment of its oyster bed. We would like to provide insights on the physiological state of the oyster if there is an infection and if there are other environmental probes we can use. We chose the oyster genes involved in response to infection as described in the litterature(7)–(9)). We chose tigriopus as an environmental probe as it was described in the litterature as accumulating v.aestearianus(10)). Finally, we chose to detect mussel genes as the Thau lagoon is also a major location of production of mussels. We also based ourselves on literature to find the targets(11)). With all these additional targets at hand, we plan to implement a multiplexed test. For this we envisioned multiple possibilities:

• Our first idea which has already been proposed in the litterature is the multiplexing through different Cas proteins(2). The nuclease activity of Cas13a although non-specific displays nucleotide preferences(4)–(6). This property of Cas13 orthologs can be used in order to multiplex. Indeed, in a system with multiple probes composed of different fluorophores each specific to a Cas-target couple we could perform a Shell’lock multiplex
• Another proposition for the multiplexing would be to physically separate the Shell’lock reaction into ‘reaction chamber’ with different guide RNAs and thus different readout.

Multiplexing targets will provide a greater amount of information in a single test. For instance, multiplexing V.aestearianus and oyster immune system targets will allow to study and correlate the dynamics of V.aestearianus spreading with actual oyster infection. These data can be extremely helpful for oyster farmers to take adapted actions based on the physiological response of their oysters.

We showed during our project that Shell’lock could be implemented to detect oyster pathogens. Throughout this section we discussed outlooks for our project and experiments to make the test more robust. We got great feedback from oyster farmers after the presentation of Shell’lock and we are looking forward to the opportunity to see it implemented on the field one day.