CADD: predicting the deleteriousness of variants throughout the human genome

By Philipp Rentzsch, Daniela M. Witten, Gregory M. Cooper, Jay Shendure, Martin Kircher. 2018

3,218 citations recorded by the source corpus.

Abstract Combined Annotation-Dependent Depletion (CADD) is a widely used measure of variant deleteriousness that can effectively prioritize causal variants in genetic analyses, particularly highly penetrant contributors to severe Mendelian disorders. CADD is an integrative annotation built from more than 60 genomic features, and can score human single nucleotide variants and short insertion and deletions anywhere in the reference assembly. CADD uses a machine learning model trained on a binary distinction between simulated de novo variants and variants that have arisen and become fixed in human populations since the split between humans and chimpanzees; the former are free of selective pressure and may thus include both neutral and deleterious alleles, while the latter are overwhelmingly neutral (or, at most, weakly deleterious) by virtue of having survived millions of years of purifying selection. Here we review the latest updates to CADD, including the most recent version, 1.4, which supports the human genome build GRCh38. We also present updates to our website that include simplified variant lookup, extended documentation, an Application Program Interface and improved mechanisms for integrating CADD scores into other tools or applications. CADD scores, software and documentation are available at https://cadd.gs.washington.edu.

Subfields: Genetics, Molecular Biology, Computational Biology & Bioinformatics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Genetics

In Molecular Biology

In Computational Biology & Bioinformatics

Citations in the corpus

Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences (2019, background): “Computational variant effect predictors are useful for assessing the effect of point mutations (Gray et al., 2018; Adzhubei et al., 2013; Kumar et al., 2009; Hecht et al., 2015; Rentzsch et al., 2018).”
Evolutionary context-integrated deep sequence modeling for protein engineering (2020, background): “The function prediction in protein engineering is also different from predicting the deleteriousness [56] or instability [42] of variants -- to assist protein engineering, the machine learning model needs to prioritize variants that are not only structurally stable and non-deleterious but also wit”
Predicting Drug Response and Synergy Using a Deep Learning Model of Human Cancer Cells (2020, background): “Alternatively, the effects of speciﬁc mutations on gene functions could be incorporated by a metric such as the Combined Annotation-Dependent Depletion score (Rentzsch et al., 2019) or by including gene structural domains as an additional layer of the hierarchy.”
dbNSFP v4: a comprehensive database of transcript-specific functional predictions and annotations for human nonsynonymous and splice-site SNVs (2020, background): “1 added 18 deleteriousness prediction scores (BayesDel_addAF and BayesDel_noAF [15], CADD_hg19 [16], ClinPred [17], DEOGEN2 [18], Eigen and Eigen PC [19], FATHMM-XF [20], GenoCanyon [21], LINSIGHT [22], LIST-S2 [23], M-CAP [24], MPC [25], MutPred [26], MVP [27], PrimateAI [28], REVEL [29], SIFT4G…”
CADD v1.7: using protein language models, regulatory CNNs and other nucleotide-level scores to improve genome-wide variant predictions (2024, methodology): “…choice of the cut-off would naturally depend on specific factors, such as the severity of the phenotype and how much expression variation is tolerated (e.g. haploinsufficiency, dominance, recessiveness), or the amount of time and resources available for curation or experimental follow-up ( 42 ).”
Ensembl 2020 (2019, background): “We have improved our annotation of non-coding variants and made Combined Annotation-Dependent Depletion (CADD; (31)) scores available for human and Genomic Evolutionary Rate Profiling (GERP) (32) scores available for all mammals.”
Estimated birth prevalence of Menkes disease and ATP7A-related disorders based on the Genome Aggregation Database (gnomAD) (2020, background): “In contrast, the REVEL and CADD ensemble approaches to assessing pathogenicity are considered far more reliable [23,24].”
Genome-wide prediction of disease variants with a deep protein language model (2022, methodology): “In contrast, computational methods that learn the biophysical properties or evolutionary constraints of protein sequences are theoretically scalable to all coding variants [15–17].”
Using deep mutational scanning to benchmark variant effect predictors and identify disease mutations (2020, neutral): “Other predictors integrate additional features including biophysical properties of amino acids, protein functional annotations and epigenetic data (Rentzsch et al , 2019).”