Bioinformatics Editing and Proofreading Services

A reviewer asks a simple question: run this again and see if you get the same answer. Eighteen months have passed, the postdoc has left, the reference genome has been updated twice, and one of the tools was installed from a branch that no longer exists. The science was sound. The record of how it was done was written as a paragraph in a methods section, and a paragraph cannot be executed.

We edit what bioinformatics groups and computational biology teams produce — methods sections and analysis descriptions, pipeline and workflow documentation, software and tool documentation for other researchers, data management and availability statements, quality control reporting and thresholds, variant calling and filtering documentation, benchmarking and validation reports, standard operating procedures for regulated analyses, data sharing and access documentation, grant and funding applications, and reports delivered to clinical or commercial clients. Our editors work on the description that has to be reproducible.

The methods and provenance description is where computational work is either reusable or lost, and it fails by describing intent rather than execution. Alignment "using standard parameters" describes a decision that was made without recording it. We work through these so every tool is named with its version and, where it matters, its source and build, since a version number is the difference between a reproducible pipeline and a story about one; so every reference and annotation resource carries its release and its date, because a variant's consequence changes between annotation releases and analyses are frequently compared across them; so non-default parameters are listed explicitly and defaults are stated as defaults for the named version, given that a tool's defaults change; so the filtering steps are given in order with the number of items remaining after each, as this single table answers most reviewer questions before they are asked; so anything done manually or ad hoc is admitted rather than absorbed into the pipeline description; and so the availability statement says where the code and the data actually are, in a form that will resolve in five years. Descriptions written this way survive the reviewer, the auditor and the person who inherits the project.

Everything you send is treated in confidence, including unpublished analyses, code and data. We are editors rather than bioinformaticians or statisticians, and we offer no view on your methods, parameters or results. What we can do is make the record complete enough to be run again.