Computational curation and analysis of publicly available protein sequence data from a single protein family

Abstract

The wealth of sequence data available on public databases is increasing at an exponential rate, and while tremendous efforts are being made to make access to these resources easier, these data can be challenging for researchers to reuse because submissions are made from numerous laboratories with different biological objectives, resulting in inconsistent naming conventions and sequence content. Researchers can manually inspect each sequence and curate a dataset by hand but automating some of these steps will reduce this burden. This paper is a step-by-step guide describing how to identify all proteins containing a specific domain with the Conserved Protein Domain Architecture Retrieval Tool, download all associated amino acid sequences from NCBI Entrez, tabulate, and clean the data. I will also describe how to extract the full taxonomic information and computationally predict some physicochemical properties of the proteins based on amino acid sequence. The resulting data are applicable to a wide range of bioinformatic analyses where publicly available data are utilized.• Step-by-step guide to gathering, cleaning, and parsing data from publicly available databases for computational analysis, plus supplementation of taxonomic data and physicochemical characteristics from sequence data.• This strategy allows for reuse of existing large-scale publicly available data for different downstream applications to answer novel biological questions.