X-twitter Youtube Linkedin Envelope

Next Event

  • DNA ME: Physics-Based, Hardware-Independent Nanopore Readout — GPU-Free Basecalling and Native-Modification Detection for Solid-State Pores and Proteins Abstract: Nanopore basecalling today means large, probabilistic neural networks trained per platform and run on GPUs. SEN takes a different route: it encodes the electrostatic environment of a molecule inside the pore and decodes it with a compact, deterministic, feed-forward model — reading sequence and native modifications (5mC / 5hmC / 6mA) GPU-free and in real time on single-digit-watt hardware (a full MinION flow cell on a Raspberry Pi 5). Because the encoding is physical rather than device-specific, SEN is a readout layer, not a sequencer: validated first on Oxford Nanopore signal at Q19, it is designed to re-parameterise to solid-state pores and to extend toward peptides, proteins and PTMs.
Date
22.07.2026
Expired!
Time
12:30 - 13:30

nanodiag BW Colloquium XXI

DNA ME

  • DNA ME: Physics-Based, Hardware-Independent Nanopore Readout — GPU-Free Basecalling and Native-Modification Detection for Solid-State Pores and Proteins Abstract: Nanopore basecalling today means large, probabilistic neural networks trained per platform and run on GPUs. SEN takes a different route: it encodes the electrostatic environment of a molecule inside the pore and decodes it with a compact, deterministic, feed-forward model — reading sequence and native modifications (5mC / 5hmC / 6mA) GPU-free and in real time on single-digit-watt hardware (a full MinION flow cell on a Raspberry Pi 5). Because the encoding is physical rather than device-specific, SEN is a readout layer, not a sequencer: validated first on Oxford Nanopore signal at Q19, it is designed to re-parameterise to solid-state pores and to extend toward peptides, proteins and PTMs.