Hi Evo 2 team,
I am using Evo 2 to analyze genomic variation in Woodhouse’s Scrub-Jay (Aphelocoma woodhouseii) with Genebank ID
GCA_048174545.1. I would like to determine whether this species was represented in the Evo 2 training data.My goal is to evaluate Evo 2 on the aphWoo1 genome as a potentially unseen genome, so I want to avoid claiming that it is out-of-distribution without confirming its absence from the training corpus.
I would appreciate if you could clarify :
1- if there were any genomic sequences from Aphelocoma woodhouseii included in Evo 2 training?
2- Also is there a species- or assembly-accession-level manifest for the NCBI eukaryotic genomes used in training that would allow users to verify whether a particular species was present?
Any clarification on how to determine species-level training-set membership would be greatly appreciated.
Thank you.
Hi Evo 2 team,
I am using Evo 2 to analyze genomic variation in Woodhouse’s Scrub-Jay (Aphelocoma woodhouseii) with Genebank ID
GCA_048174545.1. I would like to determine whether this species was represented in the Evo 2 training data.My goal is to evaluate Evo 2 on the aphWoo1 genome as a potentially unseen genome, so I want to avoid claiming that it is out-of-distribution without confirming its absence from the training corpus.
I would appreciate if you could clarify :
1- if there were any genomic sequences from Aphelocoma woodhouseii included in Evo 2 training?
2- Also is there a species- or assembly-accession-level manifest for the NCBI eukaryotic genomes used in training that would allow users to verify whether a particular species was present?
Any clarification on how to determine species-level training-set membership would be greatly appreciated.
Thank you.