TheiaProk_ONT failing raw read screening

[emmadoughty]

Many samples in the bacterial_training_data_ont dataset in https://app.terra.bio/#workspaces/theiagen-validations/Theiagen_Doughty_Sandbox/data fails raw read screening under default TheiaProk_ONT settings. NB results in the current data table have modified parameters to get assemblies.

Under default settings, genome sizes of many of the samples are massively over-estimated, making the estimated genome size much larger than the biggest known bacterial genome. I have had conversations with a couple of people, concluding that this is probably due to the lower read accuracy of ONT data leading to increased kmer discovery, which is not being considered in the genome size estimation.

Suggested fixes: Removing/replacing genome size estimation tool so that good quality data passes screening under default settings OR making recommendations to users about sequencing chemistry/basecallers to be able to run TheiaProk_ONT successfully under default settings

[sage-wright]

Closed via #164