Long Tran-Thanh – 2023-06-10 04:41:24
Below I will discuss the case of recently retracted paper of Vo et al. from the Journal of Environmental Management (2018). In particular, I will compare this paper with the 2016 paper from Nguyen et al. (Bioresource Technology). The links to these papers are below:
The 2016 paper: https://opus.lib.uts.edu.au/bitstream/10453/49966/2/Accepted%20Manuscript%20Effect%20of%20Tris-%28hydroxymethyl%29-amino%20methane%20on%20microalgae%20biomass%20growth%20%28BITE%2016093%29.pdf
The retracted 2018 paper: https://www.sciencedirect.com/science/article/pii/S0301479718305000?ref=pdf_download&fr=RR-2&rr=7d4e32a9ac4720e4
I would like to start with the summary of these papers in a layman’s language. Both papers study under which circumstances we could improve the biomass concentration of the Chlorella algae, one of the most important and widely used algae in producing biofuel. The 2016 paper investigates the how the presence of Tris (full name: tris-(hydroxymethyl)-amino-methane) affects the biomass concentration, while the 2018 paper investigates how the ratio of important nutrients (nitrogen and phosphor, or as used in the papers, the N:P ratio) affects this biomass.
Now let’s investigate the details:
The first paper’s content is clear and is technically genuine. It runs experiments (in a photobioreactor) under different settings, each run twice: one without and with the presence of Tris. The overall result shows that Tris does negatively affect the total biomass concentration of the algae. Now comes the interesting parts: the different settings here means that the authors did run the experiments under varying N:P ratio (!). In particular, they have run with N:P of 10:1, 15:1, 20:1, and 25:1, respectively. Another outcome of these experiments is that the author find out that under the 15:1 N:P ratio, the biomass concentration of the algae was the highest, with 3368 mg/L (or 105*10^6 cells/mL) – please remember these numbers (!).
The 2018 paper, on the other hand, is much more problematic. First is its title. The title indicates that the paper would vary both the nutrients’ concentrations + somehow directly manipulates the CO2 sequestration process to affect the resulting biomass concentration of the Chlorella algae. However, the paper itself does not do this, as the authors only varies the N:P ratio (sic!) and only measures the amount of sequestrated CO2 as the result of this variation. Interestingly, by only varying the N:P ratio, the authors of the 2018 paper have managed to reproduce exactly the same results of the 2016 when run without Tris (Figure 2b from the 2016 paper and Figure 2 from the 2018 paper). For example, under the 15:1 N:P ratio, the biomass concentration of the algae was the highest, with 3368 mg/L (or 105*10^6 cells/mL) – which are exactly the same values as in the 2016 paper. These results were also mentioned explicitly in text format in both papers, and one can check that they are exactly the same. The exact similarity also holds for 10:1, 20:1 and 25:1 N:P ratio values as well.
So the question is, whether the 2018 paper used the same results of the 2016 paper without rerunning the experiments (I will discuss the consequences of this later, but for now let’s just focus on its technical side).
The way both papers maintain the desired N:P ratio is to play around with the concentration of NH4Cl, K2HPO4, and KH2PO4, respectively. Table 1 in the 2016 paper and Table 1 in the 2018 paper provides information about this. At the first glance they seem to be different, so one may conclude that these are 2 different experiments. However, if you look more closely, you can see that the corresponding values in the 2018 is almost exactly half of the values in the 2016 paper. Understandably both set of values indeed result in the demanded 10:1, 15:1 etc ratios, but this raises the following question: Are these values and the results of the 2018 paper realistic?
The reason I am asking is because I believe it’s not just the nutrients’ ratio, but also their total amount affects the amount of biomass concentration Because I am no expert of the field so I cannot answer this question for sure. Instead, let me investigate the following 2 options:
1. The more realistic case is that the total amount does also affect the biomass concentration values. As the total amount of the nutrients in the 2018 are just half of that of the 2016 paper, one might expect to have different biomass concentration values as well (the trend could be the same, but the concrete values should be different). The fact that the results of the 2018 paper are identical to those of the 2016 paper raises the suspicion of research misconduct. In this case the retraction decision of the journal’s editor is absolutely justified.
1. Suppose it’s the opposite case, that is, the total amount of nutrients does not have any effects at all. Now, this should be a well-known facts in the field (which is important and let me explain later). Let’s ignore the fact that relevant literature seems to prove this is not true, for now we just accept this assumption (for the sake of complete logical investigation). In this case, the 2018 paper only differs from the first half of the 2016 paper in the measurement of CO2 sequestration. However, as the 2018 paper also explains, this amount grows linearly with the growth of biomass concentration, which is a well-known information (see section 2.6 of the retracted 2018 paper). So the fact that the 2018 paper also measured this trend in Section 3.3 does not contain any new scientific results. Note that Section 3.4 of the 2018 paper is just an application of the theory described in section 2.6 to the measured biomass. Overall, if this is the case, then the 2018 should not have been accepted in the first place (so retraction could be seen as a late fix of this mistake).
From this, we may conclude that the retraction of the paper is well justified.
Finally, let me comment on Assoc. Prof. Bui Xuan Thanh’s explanation to this case, published here:
[https://tuoitre.vn/bai-bao-quoc-te-bi-go-tac-gia-viet-nam-noi-gi-20230609223928852.htm](https://tuoitre.vn/bai-bao-quoc-te-bi-go-tac-gia-viet-nam-noi-gi-20230609223928852.htm)
While I am not in the position to judge the overall truth in his words, I think he could have done a much better job explaining the situation.
First and foremost, anh Thanh could have provided a layman’s explanation of the 2 papers so the wide audience would see what the differences are in these 2 papers (something as I have done at the beginning of this writing). By doing so, people would be able to judge whether the retraction was justified or not. I strongly believe that a good scientist, besides knowing how to do research, should be able to explain their work to the wide audience in a clear way. Without this, it’s very hard to achieve societal impact in the long term (see, e.g., all the silly recent comments and demands on realistic outputs of ALL the research in Vietnam). In fact, a good explanation would make the audience and the policy makers see the potential of the scientist’ research.
Second, he said “Việc trùng lặp một phần trong hai bài báo là do hình mô tả bể phản ứng, một số dữ liệu trong thí nghiệm. Những dữ liệu này sử dụng cùng một hệ bể phản ứng và các điều kiện thí nghiệm. Các dữ liệu này là cần thiết cho cả hai bài báo.”
Comment: From my analysis above it seems that not just the parameter settings of the environment, but the experiments themselves are also the same.
“Ngoài ra, trong bài báo thứ 2 (bị rút lại – PV), phần nội dung thảo luận cho kết quả nghiên cứu và các tài liệu tham khảo là rất khác so với bài trước đã đăng. Hệ thống kiểm tra độ trùng lặp kết quả cũng cho thấy mức độ trùng lặp xấp xỉ 2%.”
Comment: Again, apart from sections 3.3 and 3.4, the rest of the results are identical. Recall that Sections 3.3 and 3.4 are more a consequence of applying the theory to the measured biomass concentration values (which are identical to the 2016 results).
“Tuy nhiên, nhóm nghiên cứu kế thừa kết quả nghiên cứu nền tảng từ các đề tài trước, đồng thời tự đầu tư mở rộng thêm các hướng nghiên cứu khác và nhờ đó nhóm có thêm kết quả nghiên cứu mới thể hiện qua các bài báo công bố. Do đó, nhóm đã ghi lời cảm ơn quỹ trong bài báo nhưng không ghi mã số đề tài và như vậy hoàn toàn không thể dùng để nghiệm thu.”
Comment: This is a very bad habit, and one should avoid doing so. We should always be specific which grant funding supports the research. A general acknowledgement can only be included if it doesn’t come with financial support. The rationale for this is to better track the impact of invested funding in the long term. So NAFOSTED should be much stricter on this.
Statistics:
Likes: 115, Shares: 18, Comments: 7
Like Reactions: 108, Haha Reactions: 0, Wow Reactions: 1, Love Reactions: 5, Sad Reactions: 0, Angry Reactions: 0