search

yukiteruono / pbsim2

PBSIM2: a simulator for long read sequencers with a novel generative model of quality scores
GNU General Public License v2.0
69 stars 15 forks source link

how to train a new model #17

Closed yipukangda closed 9 months ago

yipukangda commented 9 months ago

HI,

If I have some raw data from a new sequencing system, how could train a specific HMM model for these data.

Thank

yukiteruono commented 9 months ago

Thank you for your using PBSIM. The building of error or quality score model is not yet fully automated. It is not yet possible to provide model building tools to users.

yipukangda commented 9 months ago

@yukiteruono I see, is there any plan for adding this function or I can train a model manually from another way. btw, do HMM model conflict with the argument --difference-ratio for error rate allocations, Thanks.

yukiteruono commented 9 months ago

There is a plan to automate and publish tools for building HMM models, but the plan is a low priority for us. The error ratio is built into error models, so --difference-ratio has no effect when using the error models. We plan to publish the model built using the latest Nanopore read on PBSIM3 in the near future.

yipukangda commented 9 months ago

@yukiteruono ok, I will follow the progression of PBSIM3, thanks.