Every scientist has to adhere to the publish-or-perish maxim to some extent – the pressure to write papers is huge, and growing. Can automated writing really help prepare manuscripts?
Scientist’s work is continuously being evaluated by how many articles you publish and where. Today, science is very competitive. There has never been so much pressure on a scientist to publish as there is today. At the same time, publishers are demanding more rigorous documentation to diminish the percentage of potential retractions and increase the reproducibility of published research.
An alarming study1 showed that out of 238 published scientific articles, barely 46% could be reproduced. Many of us are even more familiar with the recent news2 about the replication efforts which were successful for just two of five cancer papers. There is always the question of integrity, time and money.
Can we manage our time efficiently in today’s fast-paced reality – when we realise that we’re actually spending more than 40% of time on administrative tasks?
To fully understand the implications of automating paper writing, we need to take a step back and understand that in order for a software to generate a draft of a manuscript based on our own research data, it needs our data to be systematically organised and traceable.
Start with the data
The subject of automating paper writing is just the tip of the iceberg. The whole story starts with the way we do research.
Working habits differ greatly in laboratories and there is still no standardised approach to how we organise and manage research data. For example, when I was doing my PhD, my work was part of a project involving five different institutes in two different countries. I joined the project in about half way through and was also handed a lab notebook from a colleague whose work I was supposed to continue.
It turned out my coworker had the worst hand-writing and I had to chase him down a few times to get some explanations. He also used completely different logic for organising the files, so I first had to make sense of that and modify everything to fit the way I needed it.
There is barely time to think, let alone to make sure that every note, comment, document, printout or idea is neatly and systematically organised.
On the other hand, for example, I'd write changes or deviations related to my research data in my paper lab notebook and take them into account when doing data analysis. However, if someone, who wanted to learn from my work or reproduce my experiments after I left the lab, accessed the raw files without the context of the notebook, they'd inevitably get wrong results.
They would also need to invest quite some time to find all my data. I had my own space on our institute's server, where all my files would ultimately end up. Additionally, some copies were printed and added to the paper notebook and some were stored on various memory drives. I also had to share certain files with the rest of the researchers working on the project, which meant yet another copy of those files.
The integrity of my own findings depended upon the way I kept track of my data. And now, as the CEO of a software development company, I see that my work could have been done in a much more efficient way.
The situation truly is a survival mission, because the scientific community is so stretched between different rigorous work obligations, there is barely time to think, let alone to make sure that every note, comment, document, printout or idea is neatly and systematically organised.
We might say that the practice is good enough as it is. It might not seem like such a big deal and that a couple of fixes might solve most of today’s issues. But when you add them together they have a cumulative effect which, at best, can be seen as expensive time wasted on mundane tasks that could be automated and, at worst, could be responsible for the reproducibility crisis.
We need to face the fact that it is absolutely crucial that observations made in the lab get associated with electronic records as soon as possible, preferably in real time. This would contribute to better traceability of research findings and also allow labs to use the benefits of technology.
The good news is that technology is here to amplify the power of science. Not to diminish it.
Automated paper writing
The question is not whether scientists should continue taking paper notes or not – if you like it then you should continue. But labs are generating increasing amounts of digital data and many are looking for ways to improve data and project management in their labs using software.
Going digital is not easy. To be able to implement the solutions that work best for our labs, we need to change the mindset from just keeping (often scattered) records to one which defines the processes to guarantee the traceability of our work and contribute to the reproducibility of our research.
Even though all this might sound like a pile of additional work, investing a couple of weeks in defining the way you want your lab to operate and getting a software to support you every step of the way, is actually not as difficult as it might look.
Based on the conversations our team had with hundreds of scientists while conducting a study3, it turned out that electronic lab notebooks (ELNs) could be a good starting point to keep research data traceable and organised. Once that’s done, automating paper writing is the cherry on top.
The purpose of an ELN is to enable you and your team members to keep all research data in one place, organise it by projects and experiments; and communicate within the same platform. With the aim to keep it all traceable. If someone leaves the lab, all data is there. Linked, organised, understandable.
For example, when you have your results, and need feedback either from your mentor, your colleagues or external collaborators, you could simply post them a comment and they’ll be notified to respond to you. No need to send countless emails with numerous versions of the same file, chase down people or wait for meetings. For labs who collaborate from different locations, electronic lab notebooks save a lot of trouble. Activity on the project is recorded, every user has a clearly defined role and permissions and you can keep track of who did what and when. Reports on progress can be automatically generated, saving you valuable time on copy-pasting or writing. And this is just the beginning. ELN functionalities include repositories as well. All your samples, for example, can be directly assigned to your experiments and connected with generated results. This allows you to track the history of each sample.
Once all your data is organised in one place, software can gather all saved data related to the project you and your colleagues have been working on and present you with a draft of your manuscript.
Automating paper writing could save you a significant amount of time you’d spend on:
- Looking for your research data, results, notes, comments etc.
- Literature overview and looking for additional references
- Gathering up all relevant data and putting it in the form of a manuscript
- Writing the initial draft
Based on our previously mentioned study, we developed sciNote Electronic Lab Notebook. In addition to it, sciNote’s Manuscript Writer add-on that can generate a draft of a scientific manuscript based on your research data that has systematically been organized and saved in sciNote ELN, was launched in 2017.
It cannot be expected from labs to switch to digital solutions overnight. There are many things to address and coordinate.
As a scientist, I can see how stressful it is and can completely understand the labs’ need to save time wherever possible. But not at the expense of data traceability and overall integrity of your research.
To allow automated paper writing to become an accessible solution for labs, we need to address the following:
- The software can gather existing data and prepare a valid draft of the manuscript, but it cannot draw new ideas, hypotheses or conclusions out of it. It shouldn’t replace the unique mind of each scientist.
- If the software is using open access or non-open access references, they need to be referred to i.e. cited properly to avoid any risk of plagiarism.
- We need to make sure that the scientific community is aware that automatically generated texts should not be submitted for publishing without proofreading and editing.
- Steps must be taken to ensure that the software does not interfere with any ownership or other related rights in any way.
Technology is already playing an important part in the future of science by providing scientists with useful and convenient software. Aside from that, we also have the responsibility to raise awareness about the problems related to data reproducibility and time management in labs, while providing solutions available to the global scientific community.
Automating paper writing can be a benefit for many labs, but it all starts with using the benefits of technology to systematically organise your data, seamlessly collaborate with your team and partners, and efficiently manage your time.
Klemen Zupancic, PhD, is the CEO of sciNote Electronic Lab Notebook