Tips for Writing a (publishable) Scientific Article?

[The Grim Squeaker]

Howdy most hallowed, honoured (and tenured :0 ) scientist 'people'!

I wanted to ask, from anyone who's written, and more importantly, had published, any scientific papers/articles, if you have any advice to give on the subject. Absolutely anything.

I've been asked to put the results from my summer research program into an organized form, with the knowledge that we'll want to try to get it published somewhere. (Both on it's own now, though it'll also be useful as a reference for a much larger project in the future).
I've read a decent amount of articles and papers (for an undergraduate), and I know the format 'reasonably' well - (Abstract, introduction, methods+conclusions, discussion/Future directions, appendixes with supplementary material)..
However, I've never written a real, serious paper meant to be an article that has a real chance of being published. So, I wanted to ask from anyone who HAS a paper or three under their name, if you could allow me the benefit of your knowledge - from the technical level (How did you organize it - Word? Something specialized? ) to any tips on organizing the material and what/how to write. (Especially the "getting organized part").


Background - The project is in the field of Bioinformatics + Machine Learning [Biology [Field we'll be trying to publish it in probably]/applied Computer Science/Biological Chemistry]]
. It involves Prediction using the model/M.L Features I designed and coded and only CS/"theoretical" work, no "wet work" (aka physical products from a lab).

Thank you VERY Much!
(And wish me luck )

[El Moose Monstero]

In terms of what to write and how to write, that's really down to your data, whether you can use it to tell a cohesive story, and your reading of the existing literature with respect to structure and content. You should find a paper which did something similar for guidance, but only you can decide what your story is and how it fits in with the 'standard' that you find in your literature.

In terms of where to write and what to use, Word is fine - particularly if you know how to deal with the automatic formatting. Sure, you could dick around with latex or whatever it is, but there's nothing wrong with the standard Office combo IMO. Learning latex is probably a worthwhile skill in the long run, but don't get distracted screwing around with presentation tools when you could be looking at the actual content. One of the good tips I got from someone was for any graph made in Excel, always remove everything from it - any label or text, and replace it with text boxes. It gives you more control, a slightly more presentable feel. I did once hear some dick in computer science brag that they rejected any paper that had a standard Excel diagram in it. Also, if you're not already using one, a reference manager is an excellent thing to have as if you change the journal you're submitting to, it's a lot easier to reconfigure a format file than it is to reedit and reorder a bunch of references by hand.

I'd also say that you should be prepared for the fact that unless you have a very hands off supervisor, they're likely to rewrite everything (or tear the first fifteen drafts to shreds), particularly as you're an undergraduate. The only people I know who got their undergraduate projects published ending up having most of the writing done by their supervisor, even though the data and analysis were theirs. In that light, I'd make sure that 'put your results into an organised form' actually means 'write the paper' or whether it means 'give me a structured results and discussion section'.

[Ziggy Stardust]

The manner in which you write and format your paper can vary significantly depending on the journal or publication source. You should make sure to check the guidelines of whatever journal you are submitting to, and read some examples, to make sure you understand what they are looking for. For example, I am currently in the process of writing a paper about some fMRI studies I've done, and the journal in particular I am looking to submit to doesn't require a "Methods" section at all! (Instead, this particular journal wants you to put the methodology in the supplements). So it is difficult to really give any hard and fast rules when subsets of different specializations have different expectations. You should be looking for at these sorts of guidelines (all journals should have them listed on their webpage) before you begin writing.

Without knowing the specifics of your journal, the only advice I can really give is to make sure you have a "story." That is, that there is some definite and logical conclusion towards which all of the information in the paper is guiding the reader. While the "Methods" section is important, don't bog the reader down in too many details if you can avoid it. If you can outsource detailed descriptions you should do so (for example, if you used a method from an already published paper, instead of writing 3 paragraphs describing what you did, you can pare it down to 1-2 sentences with a citation). Obviously, the degree to which you want to do this is dependent on the significance of the description in question to the rest of your paper, and the degree to which you may have varied from the original methodology. Since your work is theoretical, the bulk of your writing should be focused on 1) the "problem" you are trying to "solve" (essentially, your hypothesis) and 2) comparing your theory with other published theories/methods/etc.

Word is fine, as is any other word processing software. I like to use Google Docs (or Google Drive as they now call it for some damned reason), because it makes sharing with collaborators simpler.

[The Grim Squeaker]

What do you use to manage your citations/references?
I'm trying to use Mendeley, but a lot of the articles aren't in my Mendeley library, so i'm finding it easier to just regoogle them and add them via Word's built in citation manager - but then it ends up a mess of the two. (And confusing for trying to keep track of the "citation salad").

[starslayer]

What do you use to manage your citations/references?
I'm trying to use Mendeley, but a lot of the articles aren't in my Mendeley library, so i'm finding it easier to just regoogle them and add them via Word's built in citation manager - but then it ends up a mess of the two. (And confusing for trying to keep track of the "citation salad").

Stuff like this is why I'll go against the grain and suggest you use LaTeX, since it integrates this pretty simply; it keeps track of the numbering for all your citations, figures, and sections for you. If your paper involves any math, it also makes formatting it much easier and the results usually look much better. If you don't have the time/inclination to learn it now, though - and it will take some time and work to learn - Word will get you through.

I agree with everything else Ziggy and Moose have said, especially the fact that your first drafts are going to get ripped to shreds. Also, talk to your adviser(s) about this stuff. They'll have a wealth of information and advice on what to do; I'd imagine they already have an idea of which journal/set of journals they're looking towards publishing in, the typical organization and style of papers in the field, etc.

[The Grim Squeaker]

I haven't learnt LaTex, and now that i'm done with all my statistics courses, I don't see myself doing so anytime soon. (I thought it was just for inputting math formulae digitally, not citations).

The head of my lab is - A very good researcher (tons of publications, and brilliant), but kinda disorganized. I've actually been helping her rewrite some grant proposals recently (This is Is ISrael, and my English is Awesome when not needing to actually write myself).
I REALLY hope they rip and rewrite my stuff apart (I LIKE critique and adapting vs creating), but I have the feeling they'll be leaving me to work independently as they did all throught the project. Any advisors I could talk to are in the psychobiology program, and neurobiology or psychology papers are kinda...different.
I'm also in trouble due to the project still having a lot of continued avenues - and i'm finding it difficult to "cut"the material so as to have a paper which isn't all Änd of course, our next step will be to try running with the different negative sets, etc'". (I mean, apart from the Discussion part).

Thank you very much everyone! (This is helping me sort out some of the salad in my head, after I started writing it out last night. Finished a whole draft of the ïntroduction. An incredibly crappy draft, that even I can't understand, but a section nonetheless!)

[starslayer]

LaTeX is a complete typesetting engine. Virtually every single paper you see in physics and astronomy at least is written entirely using it. On the whole, I find it makes writing papers easier. However, like I said, if you don't have the effort and time to spare right now, you don't need to use or learn it this time around.

The head of my lab is - A very good researcher (tons of publications, and brilliant), but kinda disorganized. I've actually been helping her rewrite some grant proposals recently (This is Is ISrael, and my English is Awesome when not needing to actually write myself).
I REALLY hope they rip and rewrite my stuff apart (I LIKE critique and adapting vs creating), but I have the feeling they'll be leaving me to work independently as they did all throught the project. Any advisors I could talk to are in the psychobiology program, and neurobiology or psychology papers are kinda...different.
I'm also in trouble due to the project still having a lot of continued avenues - and i'm finding it difficult to "cut"the material so as to have a paper which isn't all Änd of course, our next step will be to try running with the different negative sets, etc'". (I mean, apart from the Discussion part).

By adviser, I meant who you did your research under, not your academic adviser(s). Their name will almost certainly be on the paper, so they'll be veeery interested in what's in it and how it's written; they will also have ideas on how to pare down your work to nice little publishable nuggets. I'd be extremely surprised if they DIDN'T revise it heavily.

Really, you'll be sending drafts to everyone whose name will/might appear on the paper. They will all have their own ideas about what should be in it, its organization, the writing style, and whatnot.

[Ziggy Stardust]

I've never used LaTeX, and don't know many people that use it. I guess it just isn't as prevalent in psychology/neuroscience. I do citations the old fashioned way, which admittedly isn't very efficient, but it works for me.

I'm also in trouble due to the project still having a lot of continued avenues - and i'm finding it difficult to "cut"the material so as to have a paper which isn't all Änd of course, our next step will be to try running with the different negative sets, etc'". (I mean, apart from the Discussion part).

In my experience, the best way to try and pare things down is with a presentation. Arrange with your PI or adviser or something to give a 10-15 minute presentation of your research in front of the lab (or some other group/meeting ... I, for example, use my university's cognitive science "club" as a sounding board). Preparing something in this way really helps you get to the core of things in a way that isn't as easy with a written paper.