Scientific Writing Resources

As generative AI tools have become increasingly available to academic researchers, so too have the reports of GPT-fabricated scientific papers creeping into the public scholarly record, for example, this 2024 report from the Harvard Kennedy School:

GPT-fabricated scientific papers on Google Scholar: Key features, spread, and implications for preempting evidence manipulation | HKS Misinformation Review

Developing strong scientific writing skills has always been an important component of graduate training in the basic sciences, however, not all scientific authors have the same degree of exposure to writing classes and authorship opportunities. As the burden of recognizing fake papers is falling more and more on the readers of scientific works, there couldn’t be a better way to protect yourself against fraudulent articles than by becoming an expert at scientific writing yourself.

Here’s some resources to explore if you wish to develop your scientific writing skills:

1) E-books from the MSK Library’s collection and full-text book chapters available online

2)     Duke Graduate School Scientific Writing Resource
https://sites.duke.edu/scientificwriting/
“The Scientific Writing Resource is online course material that teaches how to write effectively. The material is not about correctness (grammar, punctuation, etc.), but about communicating what you intend to the reader. It can be used either in a science class or by individuals. It is intended for science students at the graduate level.”

“This guide to scientific writing was originally created in 2010-2011 by Nathan Sheffield for the Duke University Graduate School and funded by a Duke University Graduate School Teaching mini-grant. This current site is maintained by the Duke Graduate School. If you have questions about this site, please contact gradschool@duke.edu.”

The MSK Library also provides access to writing support tools, including:

1)     Citation Management tools – https://libguides.mskcc.org/citationmanagement  

Find out about a variety of citation management software tools that can save you time when you are formatting your manuscript’s references and bibliography.


2)     Trinka AI – https://libguides.mskcc.org/trinka

“Trinka is an AI-powered writing assistant designed for academic and technical writing. Trinka corrects advanced grammar errors and contextual spelling mistakes by providing writing suggestions in real-time. It helps academicians write in a formal, concise, and engaging manner. In addition to correcting grammatical errors, Trinka allows you to paraphrase the text and improve consistency, enabling you to enhance the quality of your writing based on your requirements.”

3)     iThenticate – https://libguides.mskcc.org/ithenticate

“iThenticate is a tool for researchers and writers to check their original works for potential plagiarism. This resource will check against 93% of Top Cited Journal content and 70+ billion current and archived web pages.” 

Questions? Ask Us at the MSK Library!

NCI’s Cancer Data Science Course

With International Love Data Week 2025 just around the corner, you might be wondering how data science could be leveraged in your own cancer research projects. Luckily, the National Cancer Institute’s Center for Biomedical Informatics & Information Technology (CBIIT) has been developing some wonderful training resources designed to help clinical oncologists and cancer researchers build their basic cancer data science skills – see https://datascience.cancer.gov/training.

Whether you have the time available to dedicate to working through a multi-chapter video course or prefer the flexibility of jumping to particular topics of interest via the online training guides, there is something useful for all types of learners with different knowledge levels.

https://datascience.cancer.gov/training/learn-data-science

https://datascience.cancer.gov/training#howcan

https://datascience.cancer.gov/training/improve-data-science-skills

NCI’s basic skills video course is a great place for beginners to start. You can work through each chapter at your own pace, watching the videos, testing your knowledge, and exploring links to extensive lists of related materials. No registration required – just jump in and start learning – gaining data science skills as you go!

https://datascience.cancer.gov/training/improve-data-science-skills/video-course/chapter/data-science-myths

Questions? Ask Us at the MSK Library!

2025 MeSH Update and PubMed Year-End Activities

At the end of each year, the National Library of Medicine (NLM) produces their annual updates to PubMed’s MeSH (Medical Subject Headings). These changes are made at every level of the MeSH infrastructure, including descriptors (headings or terms), qualifiers (subheadings), and supplementary concepts, and are made in response to changes in scientific discovery, taxonomy, ethical considerations, and published literature. The National Library of Medicine describes the necessity of these changes as this: “In biomedicine and related areas, new concepts are constantly emerging, old concepts are in a state of flux and terminology and usage are modified accordingly. To accommodate these changes, descriptors must be added to, changed or deleted from MeSH with adjustments in the related hierarchies, the Tree Structures.”

Types of MeSH Changes

  • Added Terms — brand new terms added, either as MeSH headings or Supplementary Concepts that currently do not warrant a full heading
  • Modified Terms — MeSH concepts that were changed (either name or hierarchical location), also referred to the the “Preferred Term
  • Replaced Terms — Descriptors or Supplementary concept terms are replaced by another term; this can include Supplementary Concepts upgraded to Descriptors as well as merged terms
  • Merged Terms — Multiple Descriptor or Supplementary concept terms combined under a single concept term
  • Combined Terms — Descriptor and Qualifier (subheading) combination made into a new separate concept
  • Deleted Terms — Descriptor or Supplementary concept terms removed, due to either being combined, upgraded, or renamed

What’s New in 2025

Artificial Intelligence

The 2025 Annual MeSH Update includes a variety of important and much-needed changes. One of the biggest and most needed expansions in the 2025 update is to the Artificial Intelligence concept, including dozens of new MeSH Descriptors found within the broader concept, including:

Publication types

Several changes were made to publication types for MeSH 2025, including two new publication types:

Note: The NLM has made an exception to their general rule of not retroactively indexing; so just as they did in 2019 when “Systematic Reviews” became a separate publication type, citations will be retroactively updated to reflect these two new publication types.

The Network Meta-Analysis publication type was previously a Descriptor (MeSH Heading); thus for existing citations, the Descriptor term (“Network Meta-Analysis”[MeSH]) will be replaced with either the “Network Meta-Analysis”[Publication Type] OR “Network Meta-Analysis as Topic”[Mesh] as appropriate. This is an important change, as it extends the scope of indexing the publication type back to the introduction of the original term in 2017.

The “Scoping Review”[Publication Type] will replace either the “Review”[Publication Type] or “Systematic Review”[Publication Type] on appropriate citations extending back to 2020, the first year this term appears in MEDLINE. Additionally, “Scoping Review as Topic”[Mesh] has been expanded from the “Review as a Topic”[Mesh] term.

Additionally, NLM will discontinue indexing following Publication Type terms:

Note: As with previously discontinued Publication Types, which include “Government Publication,” “Newspaper Article,” “Overall,” and “Scientific Integrity Review,” these Publication Types will continue to exist in MeSH, appear on existing citations, and be searchable in PubMed. However, they will no longer be applied to new citations.

How Does this Impact Searching?

Since there are so many changes to MeSH terms and structure, if you save search strategies or search alerts, or want to rerun a search that was previously conducted (such as from a systematic review), these changes may impact if and how your search strategy functions. Below are some steps to take to ensure that your search strategy is not only viable but also the best reflection of the current database.

  1. Check search viability — Run your search in PubMed and go to Advanced Search and check if there are issues being highlighted in the Details section. Terms that were removed or modified (name changed, upgraded, merged, etc.) will likely be in red. If your search is suddenly retrieving no results, this also could be the cause.
  2. Identify relevant new terms — Refer to the New MeSH Descriptors for 2025 list to see if there are any relevant terms that were added for 2025 that may make your search strategy more robust or specific.
  3. Identify replaced terms — Refer to the MeSH 2025 – Replace Report to identify any terms that were replaced by another term, upgraded to Descriptors, or consolidated with another term.
  4. Review publication types — Refer to the Annual MeSH Processing for 2025 NLM Bulletin for changes to publication types

If you have searches saved in your My NCBI account and/or you are getting PubMed e-mail alerts, or if you need to update your Systematic Review based on your previous search strategy, you may want to consult a Research Informationist to ensure your saved searches are not affected by the annual changes in the MeSH terminology. Don’t hesitate to ASK US!

Important Note from NLM Regarding Reindexing

Typically, the NLM does not retroactively re-index MEDLINE citations with new MeSH heading concepts. Therefore, searching PubMed for a new MeSH term tagged with [mh] or [majr] effectively limits retrieval to citations indexed after the term was introduced. Searchers may consult the MeSH database to see the previous indexing terms most likely used for a particular concept before the new MeSH heading was introduced. For terms without previous indexing information, consider the next broader term(s) in the MeSH hierarchy. or more searching guidance, see the on-demand class MeSH Changes and PubMed Searching.