nanoporetech / dorado

Oxford Nanopore's Basecaller
https://nanoporetech.com/
Other
445 stars 54 forks source link

basecalling with hac v4.2 vs v4.3 model #801

Closed eesiribloom closed 1 month ago

eesiribloom commented 1 month ago

If I have two "batches" of sequencing data and one is basecalled with: dna_r10.4.1_e8.2_400bps_hac@v4.2.0 and the other with: dna_r10.4.1_e8.2_400bps_hac@v4.3.0 Will this greatly affect my results? Should re-do basecalling so all samples are sequenced with the latest/same model? I know the most ideal situation is to have the exact same model flow cell etc. but this isn't always how things work out in real life and I dont want to waste unnecessary compute and time redoing analyses.

vellamike commented 1 month ago

Hello @eesiribloom,

v4.3 is slightly more accurate on most data, and considerably more accurate on a native DNA from a number of bacterial genomes. If your data is native bacterial DNA you may see benefits from v4.3. If not, you probably don't need to rebasecall.