Conclusions and Recommendations

This thesis set out to evaluate the suitability of Oxford Nanopore Sequencing as an analytical platform for assessing mRNA stability in comparison to capillary electrophoresis. Three specific objectives were set out:

1. Assess the stability of naked and lipid nanoparticle (LNP)-encapsulated eGFP mRNA subjected to accelerated thermal degradation studies (25°C, 35°C, and 50°C) and analyse the samples using Capillary Electrophoresis (CE) and ONS. The two methods of stability analysis were compared for their suitability for measuring mRNA stability.

2. Investigate whether mRNA fragmentation was occurring randomly or not using ONS data.

3. Identify the limitations of ONS as a tool for mRNA stability testing.

5.1. Stability Assessment: ONS vs. CE 

CE electropherograms showed a progressive loss of mRNA integrity over time across all temperatures. ONS data corroborated these findings and also allowed for the further analysis of mRNA through its sequence, coverage distributions, and fragment identification. The wealth of information gained from ONS is extremely valuable and there is likely more important information contained within the sequencing files which requires further bioinformatic analysis.

In conclusion, it may be seen from the experimental data and subsequent analysis in this thesis that ONS offers higher resolution information on mRNA stability when compared to capillary electrophoresis. However, the lack of replicates measured using ONS makes it difficult to provide a meaningful comparison between the two methods. 

5.2. ONS for Fragmentation Analysis 

Base coverage mapping and 5’ end identification revealed certain significant fragmentation “hotspots”. The presence of these same hotspots across samples supports the hypothesis that mRNA degradation is not random.

Analysis of the mRNA secondary structure revealed that some of these hotspots were located on single-stranded regions of the mRNA which are generally considered to be more susceptible to in line hydrolytic cleavage. This suggests that mRNA stability is dictated by the secondary structure of the mRNA strand. However, certain other hotspots did not follow this trend of appearing on single stranded regions meaning that further analysis of the data is required. 

5.3. ONS Limitations 

ONS’ unique ability to interrogate the sequence of every read present in the sample offers matchless resolution which makes it an important tool in the arsenal of researchers analysing mRNA stability. Unfortunately, there are several limitations associated with the method which means it is not the perfect tool for routine mRNA quality analysis. These limitations include but are not limited to: difficulty poly(A) tail profiling, inability to measure 5’ capping efficiency, inability to characterise impurities from IVT (dsRNA, DNA template, unincorporated nucleosides, RNA polymerase, etc.), the prohibitive cost of ONS, time constraints, and the need for bioinformatic expertise.

Evaluating poly(A) tail length is important as it is a CQA which affects stability and protein translation in vivo (Camperi et al., 2025). ONS has difficulty profiling poly(A) tail length because of the lack of changes in the ionic current intensity detected when sequencing homopolymer regions such as the poly(A) tail. The DRS protocol also contains a polyadenylation reaction which can interfere with the native poly(A) tail profile of the mRNA by adding an indeterminate amount of adenine residues to the 3’ end of the mRNA. This presents a major interference with the data implicit within the protocol, making the method in its current state wholly unsuitable for poly(A) tail profiling. VAX-seq, an ONS workflow developed by Gunter et al. (2023), has developed a solution to this problem by incorporating the tailfindr software into its workflow allowing for accurate and reproducible poly(A) tail estimation.

Another CQA which ONS fails to accurately measure is 5’ capping efficiency. The 5’ cap is a post transcriptional modification to the mRNA characterised by a modified guanine nucleotide (7 methylguanosine) connected to the 5’ end of the mRNA by a unique 5’-5’ triphosphate linkage. The 5’ cap affects the stability, ribosomal binding, and the translational efficiency of the mRNA transcript, clearly a CQA. Currently, the gold standard assay for 5’ cap characterisation is LC-MS/MS coupled with enzyme digestion (Hutchinson et al., 2025). Workflows have been developed in order to isolate mRNA transcripts with a 5’ cap using ONS (Jiang et al., 2019), however these have not been investigated within the context of mRNA therapy quality analysis. In our study, no meaningful information on 5’ capping efficiency could be ascertained from the ONS data.

Impurity profiling using ONS remains limited in its scope. While ONS was capable of sequencing off target mRNAs (ENO2), the detection of off target contaminants was restricted to mRNA. ONS DRS is not capable of profiling non-mRNA contaminants such as dsRNA, unincorporated nucleosides, residual RNA polymerase etc. These contaminants may produce undesired effects in vivo and detection of their presence is required in order to guarantee a product of sufficient purity is delivered to the patient.

ONS is not a particularly cost-effective measurement requiring costly reagents and equipment (flow cells) along with a complex sample preparation protocol which takes several hours. Sequencing itself takes several hours, potentially days. CE, on the other hand, is a cheap method which provides a relatively accurate and fast result. CE also allows for multiple samples (up to 15) to be run at once, while ONT’s GridION sequencer can facilitate only 5 samples at one time. The sample preparation protocol for CE using Agilent’s Tapestation 4150 is far simpler allowing for less human error.

Another obstacle preventing ONS’ widespread adoption as a quality analysis tool is the level of bioinformatics expertise required to extract meaningful information from the raw sequencing data. There is currently a lack of user-friendly tools available for the analysis of Nanopore sequencing data, which may have an impact on the implementation of ONS into mRNA quality analysis workflows when simpler alternatives, such as CE, are available.

There is an opportunity for ONS to become a widely used quality analysis tool, if used orthogonally alongside other techniques, but certain characteristics need to improve to make it a more attractive option. The most obvious improvement would be the development of user friendly reports on mRNA stability, identity, and other important quality characteristics specifically tailored to the needs of mRNA therapy manufacturers while acknowledging its shortcomings.

5.4. Future Directions

A major drawback of this study is the lack of experimental replicates used for ONS. Further experiments on the assessment of mRNA stability should be completed in the future increasing the number of replicates in order to accurately compare ONS to other methods.

The findings on mRNA degradation merits further investigation in order to better understand the mechanisms involved in mRNA stability. Investigation may be conducted on whether the inclusion of stabilising sequences in the mRNA would change the frequency of these hotspots. It may be worth investigating if the incorporation of base or backbone modifications could change the frequency of the fragmentation hotspots.

In summary, ONS provides far greater resolution than CE for mRNA stability assessment and is capable of revealing sequence level insights unattainable with other methods. However, due to its limitations with regards to cost, speed, inability to measure some of the most important mRNA quality attributes (poly(A), 5’ cap), and complex analysis, it appears that ONS is best suited as a complementary analytics tool providing orthogonal support for mRNA quality control operations. With improvements made to the supporting framework which surrounds ONS it may have a role as a powerful analytics platform within the quality control of mRNA therapeutics.