How to write scientific Reviews? Some tips from the experience of a junior researcher - a blogpost by ESR11 Hryhory Sinenka

It may be helpful!! :)

Just go to the enumerated points if you don’t have time for the backstory :)

Short backstory: as an early-stage researcher in the Allostery in Drug Discovery (ALLODD) Network within Marie Skłodowska-Curie Actions, I got lucky and honored by having an invitation by Dr. Zoe Cournia and Prof. Marco Cecchini (my Supervisor) to contribute, among many other authors, to the yet-to-be-published Review, which topic is connected with alchemical free energy calculations on certain systems, relevant to drug-design. While some final work is yet to be done before the publication, I wanted to share some of my insights on writing, which indeed was a novel experience for me.

Although at the first sight the question of “writing a Review” may seem trivial for those who at the beginning of their scientific career already have got their hands on drafting a scientific article in a peer-reviewed journal or have done decent literature work in their Bachelor’s or Master’s Thesis, writing a professional Review may require some additional techniques and persistence.

First, the difference between the typical introduction in a peer-reviewed article with your results and a Review is in how exhaustive you need to be to discuss a certain topic. Although in typical articles you certainly need to be well-informed of the current stage of scientific progress on your problem of interest, Reviews often discuss broader issues and may require a more complete literature analysis.

Additionally, in your typical article, you may know the most recent advances and key articles through your Supervisor – but a Review could go beyond the scope of their direct expertise, and includes all the most recent advances – which could’ve been missed unless a thorough work to find them was done. Just imagine how many new scientific articles are out there each month.

Thus, here are some tips for writing a scientific Review, at least what worked for me:

1) Do a preliminary search for a couple of days on the topic assigned to you. 

Use some intuitive keywords, and a familiar search engine – just to understand the current state of things in relevant questions. Read some of the recent articles and those recommended by your Supervisor and senior co-authors just to start.

2) Draft a plan for your Review with your Supervisor.

At this stage, it is just a draft that will most probably change when new information is revealed – but it will help you navigate in the ocean of information, and especially understand what you search for in the articles.

3) Think of criteria that will make the articles ineligible for your Review.

This may be anything from obvious “articles which do not use our method of interest” to really valuable separations, such as “articles, which focus on the study of protein-peptide interactions, as we are not interested in them” or “articles, which do not involve quantitative comparisons of the calculated binding affinities with the experimental values” – it depends on your focus!

4) Come up with keywords that you will use for your search, and do a sanity check.

During this important step, it may be wise to do some sanity checks: for example, you know about an important new article of 2022 from your Supervisor – will you be able to find it using your keywords? Also, see the next point, because your success may also depend on the search engine of use.

5) Try different search engines, and do a sanity check. 

This was a very valuable conclusion for me. Before I started my work on the Review, I was more than satisfied with Google Scholar during my previous junior career. But no matter what kind of keywords I used – Google Scholar failed in finding one very important article of 2022 among other articles of this year. 20 first pages, 2 different groups of keywords – no result!

Then I tried AI-powered Semantic Scholar – and voilà, I found my article on the 3rd or 4th page!

Of course, you may find many other instruments which could be helpful (you may search for something like “alternatives for Google Scholar” and so on). For example, you may try Research Rabbit app, which allows to visualize maps between the connected articles and references within them. You may find it convenient!

6) Come up with a scheme, and a protocol for how you will search for the new articles. 

For example: “I search for the articles using the keywords_1 in the Google Scholar for the first 10 pages for the year 2023, then search using the keywords_2 in the Google Scholar for the first 5 pages for the year 2023, then do the same with the other search engine, then go to the year 2022 and so on until year X”. 

Of course, you’re not a robot and can change a protocol if something seems to be too exhaustive or inefficient, and can make a protocol very detailed and flexible for different situations (for example, at some point you may say “from year X, I search all the articles for the previous years altogether for several pages and stop” and so on).

Personally, I was searching for the articles year by year, starting from the newest ones to the year 2017, because by this time many helpful reviews on the connected topics were already published. As to the older articles, I only searched them using references from the found reviews and newer works.

7) This point is connected with the previous one: decide how are you going to work with the found articles during the search. 

Are you going to read carefully the first encountered potentially interesting article and only then search for another, or at first you would like to open the links to many articles within your search engine, read their abstracts, close those not wanted, then briefly examine their texts on whether they do not contradict the criteria of point 3) and only then read the remaining carefully? These strategies are up to you. As for me, I’ve chosen the 2nd option.

Additionally, try to group the found articles by their content at some point – for example, you may group them by the type of the explored system, by their qualitative vs quantitative character, etc. It may help you to create the logic in your Review.

Remember that the Review is not just a list of articles on the same topic with the description of their content – you have to analyze them as a whole and come up with a bigger picture, connecting all the contributions into your story.

8) From time to time, discuss your draft and the changes in your planned structure of the Review with your Supervisor – especially if you have encountered a certain type of article, which eligibility for the Review is under question.

For example, be ready that the initial logic of your draft may change because of some unexpected type of articles – for example, they may explore the type of systems which you did not take into account in your initial planning!

This is it! I hope that someone finds these pieces of advice helpful and not too obvious – but of course, know that this was only the perspective of someone for whom the Review preparation was a new experience, and who worked only on certain sections of the prepared publication.