How I Write Scientific Manuscripts & Proposals

For most, scientific writing is a chore: how many people actually enjoy writing grant proposals, dealing with typesetting, or just sharing the details behind an experiment? I, for one, actually quite like writing (perhaps for the wrong reason) because it provides me an opportunity to play around with workflows, discover new command line tools, or just mess around with typography and typesetting. In this post, I’ll cover some of the steps I take when writing; be it for grants or for publications. This post will be split into three parts: working solo or collaboratively, and themes common to both.

Common themes

The principles I try and adhere to is to separate formatting from content: I will write in formats such as markdown and LaTeX, and use CSS and/or classes to define all of the nitty-gritty details like margins and section headings. The idea is to fuss minimally about formatting, which acts as my primary mode of procastination (how do I make this look prettier?) and focus up on the actual content. In this sense, I do find writing in markdown much better, although a lot of the time I will end up writing at least some LaTeX within the markdown document just to give some fine control over figure placement, etc.

I recommend deciding on a theme/format before you start depending on if you are writing in LaTeX or markdown: look at the ACS LaTeX template for the former, and the Eisvogel template for the latter as starting points.

For making figures, I nearly exclusively use Python and matplotlib. I will seldom make minor edits in Inkscape, but in the spirit of reproducibility and ease of later changes, the more you do in Python (either in a script or a notebook) the less manual and sometimes difficult to reproduce changes you have to make when some data or annotation needs to change.

I will usually use bibtex for references too: in more recent times particularly with grant proposals, I will use the biblatex LaTeX package with heavy tuning depending on the requirements of the funding agency. For example, the National Science Foundation requests two separate bibliographies, one for works supported by a previous grant, and those for the scientific justification – this is not easily done with conventional bibtex, while it is supported with the more finicky biblatex.

Another referencing note is how citations/references are formatted. In most cases it is simply changing biblatex options, but one of the largest annoyances is getting the right journal abbreviations. In my workflow, I use JabRef to clean up my .bib file of duplicates, as well as abbreviate journal names (although it could just as well be a job for sed).

Working solo

Working entirely by myself gives me complete freedom over what I can use, and how I use them:

  • Neovim for editing, along with some Vundle plugins that primarily improve quality of life
  • pdflatex or xelatex for typesetting
  • Pandoc for conversion and typesetting

While Overleaf (see next section) gives you the option of vim-like keybindings, I prefer the ability to customize them as I do locally. Similarly, working locally on Neovim lets me use plugins for formatting tables automatically, for markdown bindings, and zen-mode.

For writing letters (like those that accommpany paper submission), I much prefer using a Pandoc letter template over a Word document (shudder), whilst still keeping things like letter heads, signatures, etc. accessible.

Working collaboratively

When I write with the intention to share or work with others, Overleaf is my go-to option: everyone shares the same LaTeX environment with real-time editing; no one needs to mess with downloading and installing tex packages, and people who are more “TeX savvy” can be inclined to debug and fix compilation errors for those less proficient (particularly those messy bibliographies).

One of the great features of Overleaf is also version control on the cloud. With the ability to save and label versions of text, I usually use this feature to control iterations of manuscripts, as well as pre- and post-submission for publications as a way to keep track of changes recommended by reviewers.

Now, for some cons. As amazing as Overleaf is, it is a subscription based service and people at instutitions without coverage may find themselves missing out on most of the quality of life features, and in particular having a limited number of collaborators on a documet is somewhat restrictive. That being said, personally, some features like “Track Changes” I don’t mind missing out on, as it quickly gets visually cluttered when multiple people are editing simultaneously.


For the beginner working locally or at instutitions without a subscription to Overleaf, I highly recommend learning more about writing in markdown – it is an extremely simple syntax (basically plain text) that can be converted into many other formats. With packages like reveal-md you could even convert your documents into a presentation deck! You can always include some snippets of LaTeX for greater customization. As always, feel free to reach out to me if you have any questions, comments, or suggestions!

© 2020 Kin Long Kelvin Lee. All rights reserved.

Powered by Hydejack v8.5.1