Optimize Rawsamble for newer ONT flow cell versions (e.g., R10.4.1)

Develop optimizations of the Rawsamble pipeline to support newer Oxford Nanopore Technologies flow cell versions such as R10.4.1, adapting its indexing, filtering, seeding, and chaining parameters and/or models so that all-vs-all overlapping and de novo assembly from raw signals remain accurate and efficient on those updated signal characteristics.

Background

Rawsamble is introduced as a mechanism for all-vs-all overlapping and de novo assembly directly from raw nanopore signals, and the evaluations in this paper are conducted on ONT R9.4 flowcells. The authors explicitly state that adapting the method to newer flow cell versions is deferred.

Flow cell versions can differ in signal properties and recommended k-mer models; for example, RawHash2 (the base for Rawsamble) notes that the k-mer length used by models can vary by flow cell version. Consequently, achieving robust performance on newer versions like R10.4.1 likely requires targeted parameter and model adjustments.

References

We leave optimizations for newer flow cell versions (e.g., R10.4.1) for future work.

Rawsamble: Overlapping and Assembling Raw Nanopore Signals using a Hash-based Seeding Mechanism  (2410.17801 - Firtina et al., 2024) in Section 6 (Discussion and Future Work), Limitations