A Guide to Writing a Scientific Paper

About this guide. This guide synthesizes the FISH 290 lecture series with the assigned course readings. Where guidance differs between sources, this guide follows Mensh & Kording (2017), Ten simple rules for structuring papers (PLOS Computational Biology), which the course treats as the reference for paper structure.

Why structure matters
Four principles that govern every part of the paper
The anatomy of a paper (IMRaD + Title + Abstract)
Title
Abstract
Introduction
Methods
Results
Discussion
Acknowledgments and References
Process: how to actually write the paper
Editing: brevity, signposts, and the three-pass review
Final checklists
Sources used in this guide

1. Why structure matters

A scientific paper has one job: to make a clear claim, supported by evidence, that a reader can understand, verify, and remember. Writing is communication, so the reader's experience is the primary thing you are designing. Three audiences will read your paper, each with different priorities:

Editors — does this matter and is it sound?
Reviewers — are the conclusions justified by the data and logic?
Readers — can I quickly grasp the contribution and decide whether to dig deeper?

A well-structured manuscript serves all three by being readable, credible, and memorable. Structure isn't decoration; it is the vehicle that carries your claim across disciplinary boundaries and into your reader's long-term memory.

A paper succeeds when readers can still describe its main contribution to a colleague a year after reading it.

2. Four principles that govern every part of the paper

These four principles apply at every scale — sentence, paragraph, section, and whole paper. Get these right and the section-by-section advice that follows will feel natural.

Principle 1 — Focus on a single central contribution, and put it in the title

Be ruthlessly focused. Papers that try to make multiple separate contributions tend to be less convincing about each and are less memorable. Distill your work into one core idea and let the title carry it. The title is the first (and often only) thing most people will read; treat it as a constant reminder to you of what the paper is really about. If you ever find yourself drifting, ask: Does this sentence/paragraph/figure serve the title?

This is the "Rule of One." It is the hardest rule because science is genuinely complex, but it is the single most powerful structural decision you will make.

Principle 2 — Write for a flesh-and-blood human who does not know your work

You are the world's leading expert on what you did, which makes you the least qualified person to judge your own clarity. Adopt the perspective of a smart but uninformed reader. Concretely:

Define every technical term the first time it appears.
Avoid abbreviations and acronyms that force readers to flip back to find a definition.
Minimize the number of "loose threads" a reader has to hold in working memory. Humans remember the beginning and the end of a list far better than the middle — front-load and end-load your key information.
Avoid pronoun-heavy writing. Sentences like "It was taken to the lab" or "Temperature affected them" fail readers who do not already know what it and them refer to.

Principle 3 — Stick to Context–Content–Conclusion (C-C-C) at every scale

Memorable stories have a beginning, a middle, and an end. The same applies to scientific writing — at three nested scales:

Scale	Context (beginning)	Content (middle)	Conclusion (end)
Whole paper	Introduction	Results	Discussion
Paragraph	Topic sentence	Supporting evidence/logic	Take-home sentence
Abstract	Field & gap	What you did & found	What it means

If readers ever ask "Why was I told that?" you are missing context. If they ask "So what?" you are missing conclusion.

A common failure mode is writing chronologically — recording the order in which you actually did things. Readers do not care about your autobiography. Reformat every story into C-C-C.

Principle 4 — Optimize logical flow: avoid zig-zag, use parallelism

No zig-zag. Only the central idea should reappear repeatedly. Every other subject should be covered in one place. Related sentences belong next to each other. If you find yourself flipping back to a topic, restructure.
Use parallel form for parallel ideas. If you have three reasons to support a claim, give them the same grammatical scaffolding so the form becomes invisible and the content shines.
Don't avoid repetition for its own sake. Re-using the same noun for the same concept is good; switching synonyms makes readers wonder if you mean something different.

3. The anatomy of a paper (IMRaD + Title + Abstract)

Most scientific journals use the IMRaD format. Each section has a distinct goal and applies C-C-C in a slightly different way:

Section	Reader's question it answers	C-C-C role
Title	Is this paper worth my time?	The conclusion, compressed to one line
Abstract	What is the whole story?	A self-contained mini-paper (C + C + C)
Introduction	Why does this paper matter?	Context (sets the gap)
Methods	Can I trust how this was done?	Bridge between context and content
Results	What was found?	Content (data + logic)
Discussion	What does it mean? Where does it lead?	Conclusion (and a new, broader context)
References	Who else has worked on this?	Credit + credibility

A useful mental model from the course: the paper has five pivotal paragraphs that most readers focus on — the first paragraph of the Introduction, the last paragraph of the Introduction (the objectives), the first paragraph of the Results (the approach summary), the first paragraph of the Discussion (the answer), and the abstract. Spend disproportionate effort on these.

4. Title

The title is the most leveraged sentence in your paper. Consider the ratio of titles read to papers read — it's the difference between someone investing time in your work or never seeing it.

Properties of a strong title:

Conveys the central contribution. When in doubt, use a title that states the main message.
Tailored to the audience (which is determined by your target journal).
Right scope: too broad and you disappoint; too narrow and you undersell.
Active and informative rather than coy. "Articles with short titles describing the results are cited more often" — the title literally announces the finding.

Begin drafting your title early and revisit it often. Each revision is also a chance to sharpen what the paper is actually claiming.

5. Abstract

For most readers, the abstract is the paper. It must convey the entire message on its own. Use a tightly conserved C-C-C structure:

Structure (broad → narrow → broad)

Context (broad): One or two sentences orienting the reader to the field — comprehensible to a scientist in any discipline.
Context (narrowing): Two or three sentences of more specific background that lead the reader to the open question.
The gap: One sentence stating the specific problem this study addresses.
Content — approach: One sentence summarizing how you tackled it ("Here we show…" or its equivalent).
Content — result: Two or three sentences delivering the key findings in plain language.
Conclusion: Two or three sentences explaining what the results reveal — how they change the previous picture or extend prior knowledge.
Broader perspective (broad again): One or two sentences placing the result in a wider context, comprehensible to any scientist.

Typical length: 150–300 words. Nature's "summary paragraph" format follows essentially this same template; the example structure (with annotations) is in Nature summary paragraph on Canvas.

What does NOT belong in the abstract

In-text citations.
Detailed statistics (p-values, test statistics). Sample sizes can be included if they are easy to digest.
Vague filler ("Various results are discussed.").

Common failure mode

Talking about results before the reader is ready to interpret them. The result should "fit the gap like a key fits its lock." This nearly always requires many iterations — that is normal and expected.

6. Introduction

The job of the Introduction is to communicate why the paper matters. It is not a literature review of everything you know.

Whole-section shape: progressive narrowing toward a gap

The Introduction is a series of progressively more specific paragraphs that end with a clear exposition of what is missing from the literature, followed by a paragraph that previews how the paper fills that gap.

A canonical three-component framing from the course:

Define a research territory. The first paragraph sets the broadest scope of the paper. The breadth depends on your target journal — for a fisheries journal, start with fish or management; for a general ecology journal, start with broader ecological themes.
Establish a niche within that territory. Identify the specific gap — what is unknown that matters. Good gap statements build tension and make readers want to know the answer.
Occupy the niche. State your research question, hypotheses, and overall approach. Make it explicit why your study fills the gap you just identified.

Example progression (from broad to narrow):

¶1 — Why understanding cell differentiation matters; broadly, what triggers it is unsolved (field gap).
¶2 — Within that, what is unknown about astrocyte differentiation specifically (subfield gap).
¶3 — Clues suggest a particular gene may drive astrocytic differentiation; this hypothesis is untested (the gap you will fill).

Paragraph structure inside the Introduction

Each Introduction paragraph (except the last) should:

Orient the reader to the topic (1–2 context sentences).
Summarize what is known in the relevant literature (content).
Land on the critical unknown (the conclusion sentence — the gap at this scale).

The last paragraph of the Introduction (a pivotal paragraph)

The final paragraph is special. It:

Compactly summarizes what the paper does to fill the gap. State the objectives — clearly and, when applicable, quantitatively (e.g., "The overall goal of the study was to test the hypothesis that eelgrass density was positively correlated with fish species diversity and density. Specifically, we sampled the fish community and measured eelgrass density in eight beaches…").
Differs from the abstract in that it does not re-establish the context, may be slightly more specific about results, and previews the conclusion only briefly, if at all. Avoid spoiling the full answer here.

What to avoid

A broad literature review beyond what is needed to motivate the paper.
Tacking your project goals onto the end of an otherwise unfocused dump of background.
Awkward transitions between paragraphs that lose the reader on the way from broad to narrow.

7. Methods

The Methods section is the most procedurally straightforward to write, but it does three serious jobs:

Enable replication of the study by another competent researcher.
Provide context for the Results — every result should map onto a piece of the Methods.
Let readers judge credibility — is the approach defensible and appropriate for the question?

Structure

There is no one-size-fits-all order. Don't write strict chronology. Group like with like, in whatever order helps the reader follow the logic of the Results. Subheadings are extremely useful, and using matching subheadings in Methods and Results lets readers move back and forth easily.

A typical (but flexible) flow:

Study system / study site / organisms
Experimental design (treatments, controls, replicates)
Data collection (instruments, protocols)
Data processing / cleaning
Statistical analyses

Voice, tense, and other conventions

Past tense throughout ("We measured…", "Fish were collected…").
Active voice when possible. It is shorter, clearer, and identifies the actor:
Passive: "Six sites were selected for surveys."
Active: "We selected six sites for surveys." Use passive when the actor is genuinely irrelevant or when it makes a complex sentence clearer.
Spell out abbreviations on first use. Standard units don't need this.
Metric units, abbreviated (10.5 cm).
Military time (14:30, not 2:20 pm).
Numbers: spell out numbers that begin a sentence; otherwise, spell out numbers less than ten and use numerals for ten and above.
Scientific names: italicize genus and species (Salmo salar); after first mention, abbreviate the genus (S. salar). Higher taxa (Cichlidae) are capitalized but not italicized.
"Fish" vs "fishes": multiple individuals of one species = fish; multiple species = fishes.

"Theory of mind" while writing Methods

Constantly ask: What would a reader who is new to this work need to know here? This is the antidote to pronoun-heavy ambiguity ("It was taken to the lab…" — what is it?). Painting a clear "theatre of mind" matters most when describing:

Study sites (often accompanied by a map figure)
Experimental designs (often with a schematic)
Custom apparatus or instrumentation

8. Results (text, figures, tables)

The Results section must convince the reader that the central claim is supported by data and logic. Every scientific argument has its own logical structure, and that structure dictates the order of presentation.

How to think about a Results section

Sketch the logical chain that leads from your data to the central claim, then convert that chain into a sequence of declarative statements. These statements become your subsection headers (or figure titles, if you can't use subheaders). For example:

Method validation: our measurement is reliable.
Main effect: treatment X changes outcome Y.
Mechanism: the effect is mediated by Z.
Control: alternative explanation W is ruled out.

A clean logical progression like this makes the paper easy to follow and easy to fact-check.

The first Results paragraph (a pivotal paragraph)

Begin with a brief paragraph that summarizes the overall approach and any key innovative methods. Most readers skip the Methods section, so this paragraph gives them the gist they need.

Subsequent paragraphs follow C-C-C

Each Results paragraph should:

Set up the question the paragraph answers ("To verify that there are no artifacts…", "We next tested whether…").
Present data and logic that address it.
End with a sentence that answers the question in declarative form.

This makes paragraphs behave like little theorems: paragraphs further down rely on the conclusions of earlier ones, building toward the central claim.

What is "a result"?

Raw data (rarely shown in full in the Results — typically goes to supplements).
Summary statistics (mean, SD, n, median).
Patterns (increases, decreases, relationships).
Outputs of statistical tests (test statistic, df, p-value, r²).

The Results section primarily reports items 2–4.

Writing about results: pair the pattern with the evidence

Compare:

"The size of seeds is shown in Fig. 1." — Text doesn't stand alone.
"Fertilized seeds were larger than unfertilized seeds." — Vague; no evidence.
"Fertilized seeds were larger than unfertilized seeds (Fig. 1)." — Better, but reader must hunt.
"Mean weight of fertilized seeds was 1.8 g; unfertilized was 0.9 g." — Numbers without interpretation.
✅ "Fertilized seeds were larger than unfertilized seeds (mean 1.8 vs. 0.9 g, p = 0.019; Fig. 1)." — Pattern + supporting detail + signpost.

Other rules of the road

Past tense.
Just the facts. Save biological interpretation for the Discussion. Numerical interpretation ("Mean Hg was higher in landlocked than seagoing salmon, p = 0.02") is fine; real-world interpretation ("This suggests landlocked salmon live in more contaminated water") is not.
Use subheadings that mirror the Methods when there are multiple analyses.
Keep it short. The Results section is typically the most concise of the four body sections.
A one- or two-sentence summary of the headline findings, placed first (preferably) or last, helps readers.

Figures (the highest-leverage element in the body of the paper)

Figures are usually the only part of the body that readers who jumped in from the abstract will look at. So:

Figure titles communicate the conclusion of that figure's analysis ("Lophelia consume algae best at intermediate flow speeds").
Captions explain how the data were obtained so the figure is intelligible without the body text. Define axes, symbols, sample sizes, error bars, statistical tests, and temporal/spatial bounds. State statistical significance where relevant. Avoid interpretation in captions.
Figure captions go below figures; table captions go above tables.
Pick the right chart type for the relationship you want the reader to see (and avoid pie charts unless you have a strong reason).
The golden rule of figure design (Tufte): give readers the most information capable of being digested in the shortest time with the least ink in the smallest space possible.

When to use a figure vs. a table vs. a sentence

Sentence: few numbers, no trend. ("Copepods comprised 75% of prey items; cladocerans 12%; the remainder were unidentified.")
Figure: trends, relationships, or distributions you want the reader to see.
Table: precise numerical values, especially when there are many variables (statistical model outputs, study-site characteristics, lists of species, prior-literature comparisons).

Making tables that aren't painful

Put context columns on the left, dependent values on the right (left-to-right reading order).
Use judicious significant figures — match precision across rows. Don't paste Excel's seven decimal places.
Convention: horizontal lines above the header row and below the last row only; no internal grid lines.
Bold significant p-values so they stand out.
Every column header should be self-explanatory or defined in the caption.

Figure/table mechanics

Number figures and tables in separate series (Figure 1, Figure 2…; Table 1, Table 2…), in order of appearance in the body.
Every figure and table must be referenced in the body ("…decreased farther out (Fig. 1)" rather than "Fig. 1 shows…").
Abbreviate in text (Fig. 1) but spell out in captions (Figure 1). "Table" is never abbreviated.
Place figures and tables at the end of the manuscript, one per page (the journal will typeset them later).

9. Discussion

"Your Discussion's function is to turn data into knowledge." — Stephen Heard

The Discussion has three jobs:

Explain how the gap identified in the Introduction was filled.
Acknowledge the limitations of the interpretation honestly.
Position the contribution so that other researchers can build on it.

Whole-section shape: start specific, end broad

The Discussion is the mirror image of the Introduction. Begin by interpreting your data against your specific objectives, then progressively broaden the interpretation by bringing in other people's work, and end at roughly the same level of generality you started the Introduction with.

A workable template

Paragraph 1 — Answer the question (a pivotal paragraph). Remind the reader briefly of the question and the key results, then state — in plain language, without statistics — what your data mean. Keep it short.

Paragraphs 2–N — Detailed interpretation of each key result. Give each major finding its own paragraph. For each, ask: What does it mean biologically? Are there alternative explanations? How does it agree or disagree with prior literature? What's surprising or expected? Will it generalize?

Limitations. Set aside a paragraph (or weave limitations into the interpretation paragraphs). Never end the Discussion with limitations — the last thing the reader takes away should not be negative.

Two strategies for limitations:

"Yes, but…" — places the negative in the stress position. Avoid as the dominant frame.
"But, yes…" — "Even with a limited sample size, we detected a significant result, suggesting we had power to detect a real difference." Acknowledges the concern, then explains why the result is still credible.

When in doubt, ask yourself: What would a skeptical colleague pick at? How likely is that to materially affect the result? What evidence in the paper itself addresses it?

Broader implications and future directions. Connect back to the broad theme you opened the Introduction with. Be honest about what your results can and cannot claim. Avoid the meaningless "more research is needed"; give specific, useful breadcrumbs:

❌ "More research is needed."
✅ "Future studies might examine how temperature and pH interact during larval crab development."

Conclusion paragraph (optional but common). A short paragraph that repeats the take-home, restates how the field is now better off, and does not introduce new limitations. Some authors find this repetitive; if you use one, keep it short and end on a high note.

What to put where

Statistics belong in Results, not the Discussion. Use plain language here.
Citations belong in interpretation, limitations, and broader-implications paragraphs. Place every claim about prior literature next to a reference.

10. Acknowledgments and References

Acknowledgments

Professional courtesy. Distinct from authorship.
Recognize people who contributed (field/lab/statistical help, ideas) but whose contribution did not rise to co-authorship.
Disclose funding sources and grant numbers.
Keep it professional (this isn't a book dedication).

References (Literature Cited)

The cardinal rule: every in-text citation must appear in the reference list, and nothing else. The reference list is not a bibliography of "relevant" papers — only those actually cited.

In-text citations (APA-style baseline)

One author: Moua (2019)
Two authors: Moua and Clawson (2019)
Three or more: Moua et al. (2019) — note that "et al." is Latin, lowercase, with the period after "al."

Reference list entries (typical APA-like format)

Most journals require something close to:

Jones, R. J., and S. T. Smith. 2005. Mass mortality of long-spine sea urchins:
epidemiology and ecological consequences. Marine Biology, 74: 345–356.

Common fields: authors (last name + initials), year, full title, journal, volume, page range (or DOI/e-number for electronic-only articles).

Things to know

Every journal varies the format slightly — punctuation, ampersand vs. "and," italics, abbreviation of journal names. Always check the target journal's author guidelines (e.g., FISH 290 uses Journal of Fish Biology style).
Some journals order references alphabetically by first author's last name; others number in order of appearance in the text.
For non-journal sources, follow the journal's specific examples for books, book chapters, theses, electronic references, and pre-prints.

11. Process: how to actually write the paper

You write a good paper by allocating your effort where it matters and by iterating with feedback, not by writing it in order from start to finish.

Allocate time to what readers actually read

The title, abstract, and figures are read by far more people than the rest of the paper. The Methods section is read least of all. Budget your time accordingly.

Outline before you draft

Within each section, plan paragraphs before you write them. One useful habit: write a single informal sentence describing the role of each planned paragraph. Many writers find it helpful to start with the Results section headers — declarative statements about what each part of the data shows — because these anchor the logical chain that the rest of the paper supports.

"The point of a first draft is not to get it right, but to get it written."

Editing existing prose is much faster than generating new prose from scratch. Write something in every section, even if rough. Polished writing comes from many passes, not from one careful one.

Iterate with feedback, ruthlessly

Writing is a search through a large space of possibilities — story, outline, sentences. Sometimes a paragraph needs incremental editing; often it needs to be deleted and rewritten from a better starting point. Don't get attached to your prose.

Test readers check that the overall story works at pace.
Specific feedback (e.g., "this paragraph's logic doesn't hold") tells you where to repair the local argument.
Non-specific or unenthusiastic feedback ("I didn't quite see what you were doing") usually means the big-picture story isn't landing.
Maintain a network of helpful colleagues — and reciprocate by reading their work.

A simple test for whether you are done

If you cannot describe the entire outline of your paper to a colleague in a few minutes, your reader won't be able to either. Keep distilling.

12. Editing: brevity, signposts, and the three-pass review

Three-pass review (course-recommended)

Fine details (sentence level). Read with Track Changes on. Flag unclear pronouns, fix grammar, cut redundancy.
Overall structure (paragraph and section level). Does each paragraph serve a clear purpose? Do the sections do their jobs (see Section 13 checklist)?
Big picture (Ten Simple Rules level). Use Mensh & Kording's Table 1 (reproduced below) to diagnose what's broken at the highest level.

Brevity: every sentence must serve a purpose

Shorter writing is clearer writing.
Fluff competes for attention and gives a poor impression of the work.
Many journals and grants enforce strict length limits anyway.
A reasonable rule of thumb: cut 10–20% of your first draft.

Common redundancies to delete (drop the parenthesized words):

(already) existing
(alternative) choices
(basic) fundamentals
(completely) eliminate
(currently) underway
(empty) space
at (the present) time
never (before)
now (at this time)
period (of time)
start (out)

Common jargon → plain swaps:

Bloated	Plain
a considerable amount of	much
a considerable number of	many
a decreased amount of	less
a majority of	most
absolutely essential	essential
accounted for by the fact that	because
adjacent to	near
along the lines of	like
an adequate amount of	enough
an example of this is the fact that	for example

Signposts

Use familiar phrases to mark key statements. They reduce the reader's cognitive load.

"In this study, we tested whether…" (Introduction)
"Overall, our experiments showed that…" (Results / Discussion handoff)
"In conclusion, we found that…" (Discussion)

Reader-in-your-brain warning signs

You wrongly assumed your reader knows what you know if any of these are true:

The justification for the study is missing or weak in the Introduction.
Important context (study site, basic biology of the organism, definitions) is missing.
The writing is pronoun-heavy and the antecedents are ambiguous.

13. Final checklists

The Ten Simple Rules diagnostic table

Rule	Sign it is being violated
1. Focus on one big idea	Readers can't give a one-sentence summary of your paper.
2. Write for naïve humans	Readers don't "get" the paper.
3. Use Context–Content–Conclusion at every scale	Readers ask why something matters or what it means.
4. Optimize logical flow	Readers stumble on a small section of text.
5. Abstract = compact summary of the paper	Readers can't give an "elevator pitch" after reading it.
6. Introduction = why the paper matters	Readers show little interest in the paper.
7. Results = why the conclusion is justified	Readers don't agree with your conclusion.
8. Discussion = preempt criticism, give future impact	Readers leave with unanswered criticisms or questions.
9. Allocate time wisely	Readers struggle to grasp your central contribution despite your hard work.
10. Iterate the story	Test readers, editors, or reviewers reject the contribution.

Section-by-section structural checklist

Title
- [ ] States the central contribution (or at least its key concept).
- [ ] Calibrated to the target journal's audience.
- [ ] Not too broad, not too narrow.

Abstract
- [ ] Within 150–300 words.
- [ ] Follows broad → narrow → broad C-C-C structure.
- [ ] States the gap, the approach, the key result, and the broader meaning.
- [ ] No in-text citations or detailed statistics.
- [ ] Result actually answers the question posed in the context.

Introduction
- [ ] Starts broad and narrows progressively to the gap.
- [ ] First paragraph defines the territory; final paragraph states the objectives.
- [ ] No broad literature review beyond what's needed to motivate the study.
- [ ] Each paragraph has a topic sentence, supporting middle, and a take-home sentence.
- [ ] Objectives are explicit and, where possible, quantitative.
- [ ] Does not "spoil" the full conclusion.

Methods
- [ ] Allows the work to be replicated.
- [ ] Every Result has a matching set of Methods.
- [ ] Reader can judge the credibility of the approach.
- [ ] Grouped by topic with subheadings, not by chronology.
- [ ] Past tense throughout.
- [ ] Active voice preferred where it doesn't obscure.
- [ ] Units, time, numbers, scientific names follow conventions.
- [ ] No results disclosed here (sample sizes are OK if needed for study design).

Results
- [ ] All data needed to test every hypothesis are presented; nothing extraneous.
- [ ] No biological interpretation (save it for the Discussion).
- [ ] Each paragraph poses a question and answers it.
- [ ] Each result pairs the pattern with the supporting evidence (numbers, stats, figure call-out).
- [ ] First paragraph summarizes the approach.
- [ ] Subheadings (where used) mirror the Methods.
- [ ] Past tense throughout.

Figures and tables
- [ ] Every figure/table is necessary and not redundant with another.
- [ ] Figures show trends/relationships; tables show precise values.
- [ ] Figure titles state the conclusion; captions explain how the data were obtained.
- [ ] Every figure/table is referenced in the body and numbered in order.
- [ ] Captions go below figures and above tables.
- [ ] Each figure/table is comprehensible with its caption alone.

Discussion
- [ ] First paragraph answers the question in plain language.
- [ ] Subsequent paragraphs interpret each key result and link to prior literature.
- [ ] Limitations are addressed, but not in the final paragraph.
- [ ] "But-yes" framing used when limitations are raised.
- [ ] Concrete, specific future directions (no "more research is needed").
- [ ] Ends at the same level of generality as the Introduction's opening.
- [ ] No new statistics or new methods introduced.

Acknowledgments & References
- [ ] Funding sources and non-author contributors named.
- [ ] Every in-text citation appears in the reference list, and vice versa.
- [ ] References formatted to the target journal's style.

14. Sources used in this guide

Primary (highest priority where any conflict arises):

Mensh, B., & Kording, K. (2017). Ten simple rules for structuring papers. PLOS Computational Biology, 13(9), e1005619. readings/Ten Simple Rules Structuring Papers.pdf

Section-specific references (consistent guidance integrated):

The Introduction Section (Heard, The Scientist's Guide to Writing) — readings/The Introduction Section.pdf
The Methods Section — readings/The Methods Section.pdf
The Results Section — readings/The Results Section.pdf
The Discussion Section — readings/The Discussion Section.pdf
Dealing with Limitations — readings/Dealing with Limitations.pdf
Brevity — readings/Brevity.pdf
Scientific Writing as Storytelling — readings/Scientific Writing as Storytelling.pdf
Reading Scientific Papers — readings/Reading Scientific Papers.pdf
Nature summary paragraph — readings/Nature summary paragraph.pdf

If you remember nothing else from this guide: write for a reader who doesn't yet know your work, focus on one central contribution, organize every level of the paper as Context → Content → Conclusion, and iterate until a colleague can summarize your story back to you in a few sentences.