v5.3 release also include MisFit scores v1.5.1, a novel method to estimate missense fitness effects using a graphical model (thanks to Dr. Yufeng Shen from Columbia University, doi: 10.1038/s41467-025-59937-2).

📋 Also check dbNSFP v5.1, which included allele frequency data from 1.5 million general healthy individuals to improve rare variant interpretation:
www.dbnsfp.org/blog/highlig...

These updates provide more comprehensive and accurate variant interpretation capabilities, particularly valuable for clinical genomics applications and VUS classification.

📋 Check the detailed release notes: www.dbnsfp.org/releases

Key Updates in v5.2:
- Rebuilt variant set for all possible nsSNVs and ssSNVs based on the latest GENCODE release 48
- Added MutPred2 (thanks to Drs. Timothy Bergquist and Vikas Pejaver)
- Updated ESM1b scores (thanks to Drs. Po-Yu Lin and Nadav Brandes)
- Added GenCC for gene-disease associations

用这种方式混淆是非的人可能就是孔子所说的: 鄉愿，德之贼也。

Special thanks to Drs. Jacek Majewski, Weiva Sieh, and Ogün Adebali for updating the licenses to allow their scores to be included in the dbNSFP commercial version.

The updated score licenses are detailed in the dbNSFP README file.

As a philosopher of science, Karl Popper was not attempting to guide scientific research using philosophy (as many others did), conversely, he learned the spirit of science and used it to critique conventional philosophers and religions rooted in the pre-scientific era.

Nice sum up for the spirit of science. Ironically, I was never taught this during my time as a physics student at Fudan. Instead, I first encountered Karl Popper's concept of falsification when I read his famous book: "The Open Society and Its Enemies" right after graduation.

Thanks! Erik.

scholar.google.com/citations?us...

All these scores are readily available in dbNSFP: www.dbnsfp.org/download

Among the 60 missense variant effect predictors evaluated, MetaRNN (from dbNSFP team) and ClinPred were the top Meta-predictors on the full evaluation data set (Fig. 1) with AUROC=0.92, while AlphaMissense was the top individual predictor with AUROC=0.88.

Can you please add me to the science feed? Thanks.

google school profile scholar.google.com/citations?us...

4. Additionally, the dbNSFP team evaluated how the accumulated gene definitions improvement made in GENCODE (specifically from V46 to V47) impacts variant annotations in dbNSFP (from v5.0 to v5.1).

3. We believe this resource will not only help clinical genomic variant interpretations, but also help basic research for complex diseases such as schizophrenia and coronary artery disease, using the ultra-rare variants strategies with this much expanded reference population.

2. Some of the datasets are first-time distributed to the general public (e.g., All of Us 250K genomes) thanks to collaboration with @moezdawood.bsky.social and Dr. Richard Gibbs from HGSC at Baylor College of Medicine.