ProteinNet: a standardized data set for machine learning of protein structure

Mohammed Alquraishi

doi:10.1186/s12859-019-2932-0

Abstract

BackgroundRapid progress in deep learning has spurred its application to bioinformatics problems including protein structure prediction and design. In classic machine learning problems like computer vision, progress has been driven by standardized data sets that facilitate fair assessment of new methods and lower the barrier to entry for non-domain experts. While data sets of protein sequence and structure exist, they lack certain components critical for machine learning, including high-quality multiple sequence alignments and insulated training/validation splits that account for deep but only weakly detectable homology across protein space.ResultsWe created the ProteinNet series of data sets to provide a standardized mechanism for training and assessing data-driven models of protein sequence-structure relationships. ProteinNet integrates sequence, structure, and evolutionary information in programmatically accessible file formats tailored for machine learning frameworks. Multiple sequence alignments of all structurally characterized proteins were created using substantial high-performance computing resources. Standardized data splits were also generated to emulate the difficulty of past CASP (Critical Assessment of protein Structure Prediction) experiments by resetting protein sequence and structure space to the historical states that preceded six prior CASPs. Utilizing sensitive evolution-based distance metrics to segregate distantly related proteins, we have additionally created validation sets distinct from the official CASP sets that faithfully mimic their difficulty.ConclusionProteinNet represents a comprehensive and accessible resource for training and assessing machine-learned models of protein structure.

Highlights

Rapid progress in deep learning has spurred its application to bioinformatics problems including protein structure prediction and design
Design and approach ProteinNet’s design philosophy is simple: piggyback on the Critical Assessment of protein Structure Prediction (CASP) series of assessments to create a corresponding series of data sets in which the test set is comprised of all structures released in a given CASP, and the training set is comprised of all protein structures and sequences publicly available prior to the start date of that CASP
CASP organizers place prediction targets in two categories: template-based modeling (TBM) for proteins with clear structural homology to Protein Data Bank (PDB) entries, and free modeling (FM) for proteins containing novel folds unseen or difficult to detect in the PDB

Summary

Introduction

Rapid progress in deep learning has spurred its application to bioinformatics problems including protein structure prediction and design. In classic machine learning problems like computer vision, progress has been driven by standardized data sets that facilitate fair assessment of new methods and lower the barrier to entry for non-domain experts. While some bioinformatic applications enjoy this level of standardization [6], the central problem of protein structure prediction remains one without a standardized data set and benchmark Availability of such a data set can spur new algorithmic developments in protein bioinformatics and lower the barrier to entry for researchers from the broader machine learning community.

Methods

Results

Discussion

Conclusion