How to publish a scientific paper: Writing the paper

Before you begin

• Review the literature: Ensure that the research question has not been investigated before and that the experimental methods are appropriate. Librarians can help!
• Research data management (UC Berkeley): Ensure that your data meet disciplinary guidelines, and that you will be able to comply with funder and journal policies for data deposit and sharing.

Quick writing guides

• 11 steps to structuring a science paper editors will take seriously (Borja 2014, updated 2021)
• The Science of Science Writing (Gopen and Swan 1990)
• Short Guide to Scientific Writing (Sawyer n.d.)
• Ten simple principles for structuring papers (Mensh and Kording 2017)
• Writing workshop program
  PLOS Neglected Tropical Diseases, 2015; includes general as well as journal-specific guidelines. General guidelines begin at slide 13.

Image (detail): istock/Thomas Shanahan (Elsevier Connect)

In-depth writing guides and resources

• Nature Masterclass on Scientific Writing and Publishing
  Requires creating free Nature account, accessing from Berkeley IP address or using the Berkeley VPN with Library Access - Full Tunnel gateway, and enabling popups.
• How to Write a Good Scientific Paper (Mack 2018)
• Writing science: how to write papers that get cited and proposals that get funded (Schimel 2012)
• Graduate Writing Center (UC Berkeley)

Image (detail): Nature Masterclasses

Reporting guidelines

• Digital Curation Centre (DCC): https://www.dcc.ac.uk/guidance/standards/metadata
• Enhancing the QUAlity and Transparency Of health Research (EQUATOR): https://www.equator-network.org/reporting-guidelines/ for human- and animal-subjects research
• FAIRsharing: https://fairsharing.org/standards/

Writing tools

• Manage your citations: Citation managers EndNote, Mendeley, RefWorks, and Zotero (UC Berkeley Library guide)

• Writing + Citing in the Sciences: LaTeX editor Overleaf and citation managers (UC Berkeley Library guide)

• LaTeX in Engineering & Physical Sciences: The document formatting markup language LaTeX, which is especially useful for rendering mathematical and chemical symbols and equations (UC Berkeley Library guide)
• Virtual Online Library Tutorials (VOLT): LaTeX: Self-paced exercises for learning LaTeX

Writing tips

Avoid fragmentation (breaking a single study into multiple short papers) and redundant publication (submitting multiple papers that are very similar).

When writing a scientific paper, think about the structure familiar to you from reading scientific papers. A common structure for scientific research articles is termed IMRAD: Introduction, Methods, Results, And Discussion. A fuller outline is provided below:

1. Title: most important element; include standard, searchable terms (keywords) to call attention to your work.
   Articles with short titles describing the results are cited more often (Paiva et al. 2012); cited 200 times (Google Scholar)
    
2. Abstract: functions as an outline of the paper; include standard, searchable terms (keywords).
   1.  What is the problem domain (system under investigation)?
   2.  What is the specific research question? 
   3.  What were the methods and results? 
   4.  What are the conclusions?
   
   Example: Nunez JK, Bai L, Harrington LB, Hinder TL, Doudna JA. 2016. CRISPR immunological memory requires a host factor for specificity. Molecular Cell 62: 824-833. https://doi.org/10.1016/j.molcel.2016.04.027
   
   Abstract:
   [Problem domain] Bacteria and archaea employ adaptive immunity against foreign genetic elements using CRISPR-Cas systems. To generate immunological memory, the Cas1-Cas2 protein complex captures 30-40 base pair segments of foreign DNA and catalyzes their integration into the host genome as unique spacer sequences.  [Research question] Although spacers are inserted strictly at the A-T-rich leader end of CRISPR loci in vivo, the molecular mechanism of leader-specific spacer integration remains poorly understood.  [Methods and results] Here we show that the E. coli integration host factor (IHF) protein is required for spacer acquisition in vivo and for integration into linear DNA in vitro. IHF binds to the leader sequence and induces a sharp DNA bend, allowing the Cas1-Cas2 integrase to catalyze the first integration reaction at the leader-repeat border.  [Conclusions] Together, these results reveal that Cas1-Cas2-mediated spacer integration requires IHF-induced target DNA bending and explain the elusive role of CRISPR leader sequences during spacer acquisition. 
   
   Some journals may require a structured abstract, in which each section of the abstract (Introduction, Objectives, Methods, Results, Conclusions) is separately labelled.
    
3. Introduction: describes the general problem domain (system under investigation) and then focuses on the specific research question addressed and/or the hypothesis tested by this paper.
    
4. Methods and materials: provides enough detail to enable experiment to be reproduced by another researcher in your field. Standard experimental methods can be indicated by a reference to a published protocol.
   
   Some journals have adopted STAR Methods (Structured, Transparent, Accessible Reporting), which includes a Key Resources Table listing all reagents, antibodies, cell lines, software, or other resources required for the experiment. Use of identifiers such as Research Resource Identifiers (RRIDs) is strongly encouraged.
   
   Protocols can be shared publicly or privately on protocols.io. If they are original they can be published in a protocol journal such as bio-protocol, STAR protocols, or JoVE.
    
5. Results: describes experimental outcomes, including figures and tables.
   1. Figures: clear and compelling; each figure should tell a single story: Data Visualization Guide (UC Berkeley Library)
       
6. Discussion: explains meaning and significance of results (how do they advance the field?) and how they relate to the research question; describes limitations and further work suggested by study.
    
7. Data availability statement: Some publishers require a statement describing how the data can be accessed and reused, or the data protection concerns (such as privacy or commercial sensitivity) that prevent sharing.
    
8. Acknowledgements: Unless there are separate sections for this information, name funding sources, declare any potential competing interests, and thank contributors who are not co-authors. For human- and animal-subject research, an ethics statement may be required identifying the review committee that approved the study and the relevant guidelines and regulations that governed the research.
    
9. References: A list of sources cited in your paper. Citations (both in-text and in the reference list) must be accurate and formatted in the journal's required style. Use a citation manager.
    
10. Supplementary information: supporting technical information (figures, protocols, methods, tables, additional data) too long or detailed to fit into the body of the paper.

General tips:

• You may want to write the sections of the paper in a different order than they will appear in the published work:
  • first focus on the results (including figures and tables), discussion and methods (communicating the experimental outcomes, significance and procedures)
  • then work on the introduction, abstract and title (increasingly concise summaries of the work).
• Use simple, concrete, active language ("We determined..." not "It was determined that...")
• Start paragraphs with a topic sentence
• Consider your audience: narrowly specialized or interdisciplinary?
• Be as clear and concise as possible