A Statistical Model of Protein Sequence Similarity and Function Similarity Reveals Overly-Specific Function Predictions

Brenton Louie, Roger Higdon, Eugene Kolker · 2009

Read A Statistical Model of Protein Sequence Similarity and Function Similarity Reveals Overly-Specific Function Predictions by Brenton Louie, Roger Higdon, Eugene Kolker (2009) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0007546.txt.

Description

Brenton Louie, Roger Higdon, and Eugene Kolker published a statistical study in PLOS ONE in 2009 on predicting what proteins do. Thousands of proteins have unknown functions. Computers guess their jobs by comparing sequences to proteins already understood. One problem is that the guesses tend to be overly specific. The authors built a model relating sequence similarity to function similarity, to see where the overconfidence creeps in.

The model shows how much functional precision a given sequence match can honestly support. Beyond a point, close spelling does not justify close functional claims. Correcting that bias should make automated annotation more honest. Honest annotations matter because databases propagate guesses into thousands of later studies. The paper is a calibration tool for a workhorse method of modern biology. It tells predictors when to say less.

Catalog record

CREATORBrenton Louie, Roger Higdon, Eugene Kolker
YEAR2009
COLLECTIONPAPERS
FORMATTXT
FILE SIZE43 KB
LICENSECreative Commons Attribution License
RIGHTSCC BY
SOURCEPLOS ONE

Source: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0007546

Browse the full library · About the Tower of Babel