MetaRNN: differentiating rare pathogenic and rare benign missense SNVs and InDels using deep learning
Problem
Existing computational tools struggle to distinguish rare pathogenic from rare benign missense SNVs and non-frameshift InDels, hindering genetic analysis.
Why it matters
Prioritizing truly pathogenic variants is critical for diagnosing rare genetic diseases.
What the AI did
Deep learning models MetaRNN and MetaRNN-indel use context annotations to predict pathogenicity of rare nsSNVs and nfINDELs, with comparable scores enabling integrated genotype-phenotype association analysis.
Results
Outperforms state-of-the-art competitors on independent test sets; specific metrics not stated in source
Impact
Peer-reviewed with 190 citations; pre-computed scores and stand-alone code publicly released.
Record
| Venue | Genome Medicine |
|---|---|
| Published | 2022 |
| Work type | article |
| Authors | Chang Li, Degui Zhi, Kai Wang, Xiaoming Liu |
| Citations | 190 (OpenAlex, at import time) |
| References | 75 |
| Model or method | MetaRNN and MetaRNN-indel (deep learning) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1186/s13073-022-01120-z |
| Open access | gold |
| Keywords | MetaRNN, deep learning, pathogenicity prediction, rare variants, missense SNV, InDel |
Importance rating
significant — thousands to millions affected
Distinguishing rare pathogenic from benign variants is critical for rare genetic disease diagnosis, a foundational genetics need; abstract gives no scale numbers.