On stackoverflow the same question was asked some time ago: is LateX worth learning today? I use LaTeX, but whenever I need to share or collaborate on a text, I often have to convert to Word. I still like LaTeX because of packages like TiKZ and Sweave.
What would the answer be from a bioinformatics perspective and what are bioinformatics relevant packages?
IMHO, LaTeX is very important if you are concerned with reproducible research (in combination with Sweave and R for instance). There are of course other alternatives, like org-mode and babel for instance.
In terms of relevant packages, I remember enjoying texshade very much, to include nicely formatted sequence alignments in my documents. Although I have never used it, I remember reading about newicktree to plot phylogenetic trees. In general, I find programming my trees or networks into the documents very useful.
Regarding collaborative work, I tend to avoid Microsoft Word. Nothing can beat a concurrent versions system and raw text files (like .tex for instance), especially when there are more than a handful of authors. If LaTeX is an issue, I would invite my collaborators to get a go with Google Docs.
As a conclusion, I think that LaTeX for the sake of LaTeX is probably not essential (although advisable), but some by-products of doing things the LaTeX way, like literate programming, are crucial.
I think the best reason to learn LaTeX is that it helps you to better understand what even a WYSIWYG editor does below the surface. That helps you to understand why for instance MS-Word often behaves weird when you copy-paste things complete with formatting info (I wished the paste special as "unformatted Unicode text" option was the default).
About collaboration. Environments like Dropbox allow for a reasonable version and collision control system that even Biologists can use. And it is probably easier for you to learn Word than for the average Biologist to learn LaTeX. Of course you can embed LaTeX formatted objects if that is important or convenient.
My documents are normally a makefile that runs a variety of xelatex, bibtex, and gnuplot on input files in order to generate the required output, and with everything in version control. But this isn't a solution for everyone, and normally, bioinformaticians will collaborate with biologists (who only will know Word), and the real typesetting will be performed by journals, so you don't need to care much about output quality. So, no, I don't think you need to learn LaTeX. But maybe you should anyway?