# Creating a scientific manuscript in LaTeX

I’m submitting a manuscript as a .tex file for the first time, so I thought I’d document my workflow for anyone else who wants to give it a try. This is meant for people who know the very basics of LaTeX but are essentially beginners (much like myself.)

I habitually write my manuscripts in Scrivener, which allows me to organize my thoughts into the structure of a scientific paper.

Scrivener has an option to compile drafts to LaTeX via MultiMarkdown. The export is not perfect, but it doesn’t take too much tweaking.

I got started with Elsevier’s article template, even though I am not submitting to an Elsevier journal, and customized from there. One of the first customizations was to find the right .bst file, which BibTeX uses to format the bibliography according to a particular journal’s style. Then I exported the Zotero collection with all my citations for the manuscript to a .bib file and put it in the folder with my .tex file.

Then I had to go through the manuscript and replace all the parenthetical citations I’d written by hand and replace them with natbib syntax. I cited R using the syntax \citep{r2018}, where r2018 was my keyword for R in the .bib file with all my citations:

@misc{r2018,
title = {R: A Language and Environment for Statistical Computing},
author = {{R Core Team}},
organization = {R Foundation for Statistical Computing},
year = {2018},
url = {https://www.R-project.org/},
}

We did our analyses in R \citep{r2018}.

I also had to add captions and labels to my figures so I could refer to them in the text with their labels instead of their numbers. The advantage to doing this is that if I were to delete Figure 1 in a later revision, all the references to figures would automatically update to reflect the fact that the numbering of the figures has now changed.

\begin{figure}[p!]
\includegraphics[keepaspectratio,width=\textwidth,height=0.75\textheight]{somefigure.png}
\label{fig:somefig}
\end{figure}

Our methods were very clever (see Figure \ref{fig:somefig}).

The journal I am submitting to requires that every figure have its own page in the manuscript. The [p!] modifier after \begin{figure} ensures this.

Fortunately, I had already made my tables in LaTeX, so I didn’t have to convert those over. However, if I hadn’t already made my tables in LaTeX, I could have made them using the handy R package xtable.

library(xtable)
school <- data.frame(Year = c("2008", "2013", "2018"),
Schooling I Had = c("High School", "College", "Masters"))
xtable(school, c("Level of schooling acquired by JAP"), label = "table:school")
## % latex table generated in R 3.5.1 by xtable 1.8-3 package
## % Fri Nov 30 16:15:04 2018
## \begin{table}[ht]
## \centering
## \begin{tabular}{rll}
##   \hline
##  & Year & Schooling.I.Had \\
##   \hline
## 1 & 2008 & High School \\
##   2 & 2013 & College \\
##   3 & 2018 & Masters \\
##    \hline
## \end{tabular}
## \caption{Level of schooling acquired by JAP}
## \label{table:school}
## \end{table}

I employed several packages to make my manuscript conform to the journal’s requirements.

\usepackage{graphics}
\usepackage{hyperref}
\usepackage{mathptmx}
\usepackage{setspace}
\doublespacing
\usepackage[margin=1in]{geometry}
\usepackage{lineno}

graphics is the basic graphics package for embedding figures in the manuscript. My manuscript has a hyperlink to a website, which is enabled by hyperref. I was required to use Times New Roman font, which is enabled by mathptmx. setspace makes it easy to set the line spacing in the document to 1.5, double, or a custom spacing. geometry lets you set the size of the margins, and lineno adds line numbering.

I’m considering writing my manuscript in LaTeX from the beginning, next time. The only obstacle to that is the review process. It is much easier to comment on a document in Microsoft Word than it is to comment up a PDF, as my advisor pointed out to me when I sent him a PDF of my latest draft. I have yet to find a solution to this, so for now I’m converting my .tex file to .docx using pandoc. This is what that bash script looks like:

pandoc --filter pandoc-citeproc -o manuscript.docx manuscript.tex

This process is imperfect, especially when it comes to tables. They just don’t get translated at all. I had to open up the Word doc and insert them by hand, which was a nuisance, so I do hope to find a better workflow for the feedback cycle of a scientific manuscript. Otherwise, I found the total control I had over the presentation of my manuscript in LaTeX to be very satisfying.