A framework for exhaustively identifying functional missense variation applied to four human proteins
Ontology highlight
ABSTRACT: Four deep mutational scanning maps for human genes: UBE2I (Sumo E2 conjugase), SUMO1 (Sumo1), TPK1 (Thiaminepyrophosphokinase 1) and CALM1 (Calmodulin)
The files contain the results of Deep Mutational Scans, followed by imputation of missing values using machine learning and refinement of less confidently measured values
There are two files for each file. One with original scores, and one with hypercomplementing variant scores inverted ("flipped") to reflect deleteriousness
Score scaling: 0 = nonsense-like; 1 = synonymous-like
Columns:
mut: variant descriptor wt-residue followed by amino acid position, followed by mutant residue
screen.score: Experimentally measured score (NA where missing)
screen.sd: Standard deviation of the experimentally measured score (NA where missing)
df: Degrees of freedom / number of measurements underlying the score
predicted.score: Random Forest-based prediction for the score
joint.score: "Refined" or "Imputed" score.
joint.sd: Standard deviation of refined scores (or RMSD of imputed score, where applicable)
joint.se: Standard error of the mean for refined scores (or RMSD of imputed score, where applicable)
SUBMITTER: Jochen Weile
PROVIDER: S-BSST60 | biostudies-other |
REPOSITORIES: biostudies-other
ACCESS DATA