Model
DNABERT generates a dense representation of a genome sequence by identifying contextually similar sequences in the human genome.
Sign in to access this content
| Field | Response |
|---|---|
| Participation considerations from adversely impacted groups (protected classes) in model design and testing: | Gender |
| Measures taken to mitigate against unwanted bias: | Chromosome 21 and 22 were held out to assess intra-chromosomal bias captured by the model. No training was performed on chromosomes 21 and 22 to preserve a completely independent sequence for testing. Additionally, training was only performed on autosomes as a gender proxy (no X/Y chromosomes) to mitigate bias introduced by sex-specific DNA sequences. |