Introduction
Aquaculture plays a vital role in the food supply, especially as fishery resources decline and marine pollution increases. It provides high-quality protein and helps reduce the exploitation of marine resources, promoting sustainability. However, bacterial contamination, particularly from A. hydrophila, poses a major challenge. This bacterium uses quorum sensing, via C4-HSL, to produce virulence factors that cause infections, reducing aquaculture yields. Traditional chemical sterilization methods, though effective, often leave behind harmful chemical residues.
Therefore, we considered modifying B. subtilis WB600 by genetic engineering to express 6 different AHL lactonase genes, which enabled it to specifically degrade C4-HSL. Thus blocking the pathway of A. hydrophila virulence factor production.
The modeling part was divided into two parts:
Model 1: Hypothesis testing for the validity
To verify the validity of the bacteria-inhibiting regimen, we applied a hypothesis test. Data were collected by setting up a control group (an unmodified strain and a commercial strain) and experiment groups (six different expressing strains). Hypothesis testing helped us analyze the differences in inhibitory effects between different expressing strains. This model directly guides which expressed strains are ultimately used in our products.
1. Descriptive analysis
According to the experimental design, we perform five experiments: the zone of inhibition test, the synthetic AHL degradation test, the natural AHL degradation test, the biofilm reduction test, and the extracellular protease reduction test (for more detailed information, see https://2024.igem.wiki/hangzhou-biox/engineering).
Table 1: The zone of inhibition test
|
A.h. |
WB600 |
pHT43 |
BsAiiA |
BsYtnP |
MtAiiM |
AtAttM |
AsAhlD |
SsAhlS |
Commercial bacteria |
|
|
Antibacterial circle diameter mm |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
Table 1 suggests that none of the groups has a significant antibacterial area. It implies that these strains cannot inhibit bacteria by killing bacteria. Actually, our method is to genetically engineer B. subtilis to degrade C4-HSL, thus preventing A. hydrophila from generating virulence factors. Consequently, the genetically engineered B. subtilis itself does not directly inhibit bacteria. It just interferes with the production of virulence aspects through quorum sensing, then restricting the growth of bacteria.
Table 2: Synthetic AHL degradation test
|
WB600 |
pHT43 |
BsAiiA |
BsYtnP |
MtAiiM |
AtAttM |
AsAhlD |
SsAhlS |
BsAiiA+BsYtnP |
BsAiiA+MtAiiM |
BsYtnP+MtAiiM |
Commercial Strain |
|
|
Colorless ring diameter mm |
11.5 |
11.5 |
21.5 |
17.5 |
16.5 |
13 |
16 |
16 |
20.5 |
19 |
18 |
12 |
|
Repeat 2 |
12.5 |
12 |
22.5 |
18 |
17.5 |
12.5 |
17 |
17 |
21 |
19 |
18.5 |
13.5 |
|
Repeat 3 |
12 |
13 |
22 |
18 |
17 |
12.5 |
15.5 |
16.5 |
21 |
17.5 |
18 |
13 |
|
Colorless radius mm |
1.5 |
1.5 |
6.5 |
4.5 |
4 |
2.25 |
3.75 |
3.75 |
6 |
5.25 |
4.75 |
1.75 |
|
Repeat 2 |
2 |
1.75 |
7 |
4.75 |
4.5 |
2 |
4.25 |
4.25 |
6.25 |
5.25 |
5 |
2.5 |
|
Repeat 3 |
1.75 |
2.25 |
6.75 |
4.75 |
4.25 |
2 |
3.5 |
4 |
6.25 |
4.5 |
4.75 |
2.25 |
|
Mean colorless radius |
1.8 |
1.8 |
6.8 |
4.7 |
4.3 |
2.1 |
3.8 |
4.0 |
6.2 |
5.0 |
4.8 |
2.2 |
|
p-value |
0.76764 |
0.00002 |
0.00006 |
0.00026 |
0.11612 |
0.00138 |
0.00039 |
0.00001 |
0.00035 |
0.00005 |
0.18900 |
We examined the diameter generated by different strains of B. subtilis or combinations. We treated WB600 as control group 1 while setting the pHT43 as the control group 2. BsYtnP has the largest diameter, suggesting high degradation activity, while AtAttM produced the smallest one. Compared with the control groups, a larger diameter in the experiment groups usually means higher AHL lactonase activity in it.
Table 3: Natural AHL degradation test
|
A.h. |
WB600 |
pHT43 |
BsAiiA |
BsYtnP |
MtAiiM |
AtAttM |
AsAhlD |
SsAhlS |
Commercial bacteria |
|
|
Colorless ring diameter mm |
24 |
24.5 |
24 |
/ |
/ |
/ |
24 |
/ |
/ |
24 |
|
Repeat 2 |
27 |
25 |
26 |
/ |
/ |
/ |
25 |
/ |
/ |
22.5 |
|
Repeat 3 |
25.5 |
24 |
24.5 |
/ |
/ |
/ |
24.5 |
/ |
/ |
25.5 |
|
Colorless radius mm |
7.75 |
8 |
7.75 |
0 |
0 |
0 |
7.75 |
0 |
0 |
7.75 |
|
Repeat 2 |
9.25 |
8.25 |
8.75 |
0 |
0 |
0 |
8.25 |
0 |
0 |
7 |
|
Repeat 3 |
8.5 |
7.75 |
8 |
0 |
0 |
0 |
8 |
0 |
0 |
8.5 |
|
Mean colorless radius |
8.5 |
8.0 |
8.2 |
0.0 |
0.0 |
0.0 |
8.0 |
0.0 |
0.0 |
7.8 |
Table 3 showed that strains AtAttM, pHT43, and WB600 formed clear tinted circles, showing very low AHL lactonase activities, unlike strains BsAiiA, BsYtnP, MtAiiM, SsAhlS, and AsAhlD, which revealed high activity. We utilized the mean zone diameter to more accurately stand for the AHL lactonase activity levels.
Table 4: Biofilm reduction test
|
WB600 |
pHT43 |
BsAiiA |
BsYtnP |
MtAiiM |
AtAttM |
AsAhlD |
SsAhlS |
BsAiiA+BsYtnP |
BsAiiA+MtAiiM |
BsYtnP+MtAiiM |
Commercial bacteria |
A.h. |
|
|
Absorbance OD570/OD600 |
0.2539 |
0.2006 |
0.0334 |
0.0442 |
0.0334 |
0.2309 |
0.012 |
0.0462 |
0.0334 |
0.0465 |
0.0442 |
0.1948 |
0.2601 |
|
Repeat 2 |
0.2523 |
0.2151 |
0.0458 |
0.0867 |
0.0442 |
0.1081 |
0.0423 |
0.06 |
0.0442 |
0.0423 |
0.012 |
0.247 |
0.2631 |
|
Repeat 3 |
0.0692 |
0.1992 |
0.0258 |
0.0637 |
0.0653 |
0.2042 |
0.083 |
0.0472 |
0.0281 |
0.0334 |
0.0495 |
0.1385 |
0.2089 |
|
Repeat 4 |
0.247 |
0.2737 |
0.0403 |
0.037 |
0.0761 |
0.2684 |
0.0867 |
0.046 |
0.0495 |
0.0547 |
0.0745 |
0.2203 |
0.2429 |
|
Repeat 5 |
0.199 |
0.2187 |
0.0423 |
0.0281 |
0.0067 |
0.2364 |
0.0265 |
0.0656 |
0.0814 |
0.0173 |
0.0708 |
0.1619 |
0.2723 |
|
average |
0.2043 |
0.2215 |
0.0375 |
0.0519 |
0.0451 |
0.2096 |
0.0501 |
0.053 |
0.0473 |
0.0388 |
0.0502 |
0.1925 |
0.2495 |
|
Average reduction |
18.1% |
11.2% |
85.0% |
79.2% |
81.9% |
16.0% |
79.9% |
78.8% |
81.0% |
84.4% |
79.9% |
22.8% |
|
|
p-value |
0.25679 |
0.15092 |
0.00000 |
0.00000 |
0.00000 |
0.21437 |
0.00001 |
0.00000 |
0.00000 |
0.00000 |
0.00000 |
0.03507 |
Table 4 summarizes the effects of different experimental groups on biofilm formation. Biofilms are complex microbial communities formed by microorganisms adhering to surfaces during growth. They are highly resistant to antibiotics and the host immune system, making microbial infections difficult to treat.
The absorbance value with WB600 has a reduction rate of 18.1%, though the p-value was above 0.05, indicating that while there was increased biofilm development, it was not statistically significant. However, the large absorbance suggests that bacteria can still proliferate easily in this environment. In contrast, the mean reduction rate in other experimental groups was higher, and the p-value was significantly lower than 0.05, indicating effective biofilm inhibition. Specifically, pHT43, BsYtnP, MtAiiM, and AtAttM demonstrated significant inhibition. Although SsAhlS showed a reduction rate of 22.8%, it was lower than the more effective groups, despite a p-value well below 0.05. This suggests that while SsAhlS has an inhibitory effect on biofilm formation, its performance is weaker compared to the other groups.
Table 5: Extracellular protease reduction test
|
B.s. fermentation broth +A.h. |
WB600 |
pHT43 |
BsAiiA |
BsYtnP |
MtAiiM |
AtAttM |
AsAhlD |
SsAhlS |
BsAiiA+BsYtnP |
BsAiiA+MtAiiM |
BsYtnP+MtAiiM |
Commercial bacteria |
|
Enzyme activity (U/mL) |
30.56 |
36.36 |
20.01 |
16.93 |
22.18 |
32.41 |
22.06 |
22.5 |
20.17 |
24.75 |
22.82 |
/ |
|
Repeat 2 |
40.95 |
38.16 |
17.08 |
24.06 |
21.59 |
36.06 |
22.99 |
24.43 |
20.01 |
18.04 |
21.65 |
/ |
|
Repeat 3 |
31.2 |
35.58 |
21.5 |
23.44 |
20.36 |
36.63 |
24.1 |
20.37 |
21.39 |
20.41 |
19.54 |
/ |
|
average |
34.24 |
36.7 |
19.53 |
21.48 |
21.38 |
35.03 |
23.05 |
22.43 |
20.52 |
21.07 |
21.34 |
/ |
|
Average reduction |
-7.18% |
42.96% |
37.27% |
37.56% |
-2.31% |
32.68% |
34.49% |
40.07% |
38.46% |
37.68% |
/ |
|
|
p-value |
0.5144 |
0.0151 |
0.0348 |
0.0195 |
0.8362 |
0.0306 |
0.0295 |
0.0155 |
0.0277 |
0.0210 |
/ |
Table 5 shows the impact of various B. subtilis strains on the protease activity of A. hydrophila. We set B. subtilis culture supernatant + A. hydrophila, with a protease activity of 34.24 U/mL as the standard. The protease activity value reflected the strain's impacts on protease activity. Compared to the baseline, the WB600 revealed a rise of 7.18%, while the BsAiiA showed a 42.96% decrease. This shows that the non-genetically changed WB600 team hardly weakened proteases, whereas the genetically modified BsAiiA efficiently degraded them.
2. Hypothesis Testing
We outline the hypotheses for experiments as follows:
2.1 Synthetic AHL degradation test
Null Hypothesis (H0): There are no differences in C4-HSL degradation between experimental and control groups.
Alternative Hypothesis (H1): There is a difference in C4-HSL degradation.
2.2 Natural AHL degradation test
Null Hypothesis (H0): There are no differences in the degradation of A. hydrophila-produced C4-HSL between experimental and control groups.
Alternative Hypothesis (H1): There is a difference in the degradation of A. hydrophila-produced C4-HSL.
2.3 Biofilm reduction test
Null Hypothesis (H0): There are no differences in biofilm inhibition between experimental and control groups against A. hydrophila.
Alternative Hypothesis (H1): The experimental group inhibits A. hydrophila biofilm formation better than the control group.
2.4 Extracellular protease reduction test
Null Hypothesis (H0): There are no differences in protease activity reduction between experimental and control groups.
Alternative Hypothesis (H1): The experimental group decreases protease activity more than the control group.
3. Method Selection
Considering that experimental data are repeatedly measured, and experimental groups has several categories. Thus, ANOVA (Analysis of Variance) is perfectly suitable for testing differences. Nonetheless, executing ANOVA must satisfy its assumptions, such as normality, homogeneity of variance, and independence. If it is not met, we utilize the Kruskal-Wallis H test. If the ANOVA works well and shows considerable differences, Tukey's Honestly Significant Difference (HSD) test for multiple comparisons will certainly be carried out to determine the source of distinctions. We can also use the following criteria to select appropriate statistical methods.
4. Results Analysis
4.1 Synthetic AHL degradation test
Normality
We conducted the Shapiro-Wilk test to assess the assumption of normality. The data for WB600, pHT43, BsAiiA, MtAiiM, AsAhlD, and SsAhlS do not significantly deviate from a normal distribution (P-value >= 0.05). And, the data for BsYtnP, AtAttM, BsAiiA+ BsYtnP, BsAiiA+ MtAiiM, and BsYtnP+ MtAiiM significantly deviate from a normal distribution (P-value <= 0.05).
Table 6: Shapiro-Wilk test
|
Group |
Shapiro-Wilk statistics |
P-value |
Normality |
|
WB600 |
1.0 |
1.0 |
yes |
|
pHT43 |
0.964 |
0.637 |
yes |
|
BsAiiA |
1.0 |
1.0 |
yes |
|
BsYtnP |
0.750 |
7.772e-16 |
no |
|
MtAiiM |
1.0 |
1.0 |
yes |
|
AtAttM |
0.750 |
7.772e-16 |
no |
|
AsAhlD |
0.964 |
0.637 |
yes |
|
SsAhlS |
1.0 |
1.0 |
yes |
|
BsAiiA+BsYtnP |
0.750 |
7.772e-16 |
no |
|
BsAiiA+MtAiiM |
0.750 |
7.772e-16 |
no |
|
BsYtnP+MtAiiM |
0.750 |
0.0 |
no |
|
Commercial |
0.964 |
0.637 |
yes |
Homogeneity of Variance Test
We used Levene's test to check the homogeneity of variance. The Levene statistic is 0.021304067140090345, indicating a small difference between groups. The p-value was 0.9789347135647807, which was much higher than 0.05. Thus, we conclude that there is no significant difference in variance between the groups.
Table 7: Levene Test
|
Group |
Levene statistics |
P-value |
Homogeneity of variance judgment |
|
All groups |
0.021304067140090345 |
0.9789347135647807 |
satisfy |
Kruskal-Wallis H Test
The Kruskal-Wallis H test statistic is 33.261, which is relatively large. The P-value is 0.000477, which is much smaller than 0.05. This means that the observed data distribution strongly contradicts the original assumption that all sample groups come from the same population. Therefore, we reject the null hypothesis and conclude that the distribution of at least one sample group is different from the others. This suggests that at least one of the multiple sample groups has a significantly different distribution than the others.
Then, we need to do Post-Hoc Multiple Comparisons
Table 8: Kruskal-Wallis H Tests
|
Group |
H statistic |
P-value |
Test result |
|
WB600, pHT43, BsAiiA, ... Commercial |
33.261 |
0.000477 |
reject the null hypothesis that at least one sample has a different distribution from the others |
Post-Hoc Multiple Comparisons (Mann-Whitney U Test)
We will use the Mann-Whitney U test (also known as the Wilcoxon rank sum test) to make pound-for-pair comparisons between each pair of sample groups. It can help identify which specific groups show significant differences. Note that during multiple comparisons, we will adjust the p-value to control the error detection rate (using Bonferroni correction).
Table 9: Mann-Whitney test
|
Group1 |
Group2 |
U-statistic |
p-value |
Adjusted p-value |
Group1 |
Group2 |
U-statistic |
p-value |
Adjusted p-value |
|
WB600 |
pHT43 |
4.000 |
1.000000 |
66.000000 |
pHT43 |
BsAiiA |
0.000 |
0.100000 |
6.600000 |
|
WB600 |
BsAiiA |
0.000 |
0.100000 |
6.600000 |
pHT43 |
BsYtnP |
0.000 |
0.076523 |
5.050485 |
|
WB600 |
BsYtnP |
0.000 |
0.076523 |
5.050485 |
pHT43 |
MtAiiM |
0.000 |
0.100000 |
6.600000 |
|
WB600 |
MtAiiM |
0.000 |
0.100000 |
6.600000 |
pHT43 |
AtAttM |
2.500 |
0.500184 |
33.012161 |
|
WB600 |
AtAttM |
1.000 |
0.164160 |
10.834542 |
pHT43 |
AsAhlD |
0.000 |
0.100000 |
6.600000 |
|
WB600 |
AsAhlD |
0.000 |
0.100000 |
6.600000 |
pHT43 |
SsAhlS |
0.000 |
0.100000 |
6.600000 |
|
WB600 |
SsAhlS |
0.000 |
0.100000 |
6.600000 |
pHT43 |
BsAiiA+BsYtnP |
0.000 |
0.076523 |
5.050485 |
|
WB600 |
BsAiiA+BsYtnP |
0.000 |
0.076523 |
5.050485 |
pHT43 |
BsAiiA+MtAiiM |
0.000 |
0.076523 |
5.050485 |
|
WB600 |
BsAiiA+MtAiiM |
0.000 |
0.076523 |
5.050485 |
pHT43 |
BsYtnP+MtAiiM |
0.000 |
0.076523 |
5.050485 |
|
WB600 |
BsYtnP+MtAiiM |
0.000 |
0.076523 |
5.050485 |
pHT43 |
Commercial |
2.000 |
0.368688 |
24.333426 |
|
WB600 |
Commercial |
1.500 |
0.268286 |
17.706868 |
BsYtnP |
MtAiiM |
8.500 |
0.115688 |
7.635409 |
|
BsAiiA |
BsYtnP |
9.000 |
0.076523 |
5.050485 |
BsYtnP |
AtAttM |
9.000 |
0.072198 |
4.765081 |
|
BsAiiA |
MtAiiM |
9.000 |
0.100000 |
6.600000 |
BsYtnP |
AsAhlD |
9.000 |
0.076523 |
5.050485 |
|
BsAiiA |
AtAttM |
9.000 |
0.076523 |
5.050485 |
BsYtnP |
SsAhlS |
9.000 |
0.076523 |
5.050485 |
|
BsAiiA |
AsAhlD |
9.000 |
0.100000 |
6.600000 |
BsYtnP |
BsAiiA+BsYtnP |
0.000 |
0.072198 |
4.765081 |
|
BsAiiA |
SsAhlS |
9.000 |
0.100000 |
6.600000 |
BsYtnP |
BsAiiA+MtAiiM |
2.500 |
0.493563 |
32.575144 |
|
BsAiiA |
BsAiiA+BsYtnP |
9.000 |
0.076523 |
5.050485 |
BsYtnP |
BsYtnP+MtAiiM |
2.000 |
0.301700 |
19.912172 |
|
BsAiiA |
BsAiiA+MtAiiM |
9.000 |
0.076523 |
5.050485 |
BsYtnP |
Commerial |
9.000 |
0.076523 |
5.050485 |
|
BsAiiA |
BsYtnP+MtAiiM |
9.000 |
0.076523 |
5.050485 |
AtAttM |
AsAhlD |
0.000 |
0.076523 |
5.050485 |
|
BsAiiA |
Commercial |
9.000 |
0.100000 |
6.600000 |
AtAttM |
SsAhlS |
0.000 |
0.076523 |
5.050485 |
|
MtAiiM |
AtAttM |
9.000 |
0.076523 |
5.050485 |
AtAttM |
BsAiiA+BsYtnP |
0.000 |
0.072198 |
4.765081 |
|
MtAiiM |
AsAhlD |
7.500 |
0.268286 |
17.706868 |
AtAttM |
BsAiiA+MtAiiM |
0.000 |
0.072198 |
4.765081 |
|
MtAiiM |
SsAhlS |
7.000 |
0.368688 |
24.333426 |
AtAttM |
BsYtnP+MtAiiM |
0.000 |
0.072198 |
4.765081 |
|
MtAiiM |
BsAiiA+BsYtnP |
0.000 |
0.076523 |
5.050485 |
AtAttM |
Commercial |
3.500 |
0.822187 |
54.264327 |
|
MtAiiM |
BsAiiA+MtAiiM |
0.500 |
0.115688 |
7.635409 |
AsAhlD |
SsAhlS |
3.000 |
0.653095 |
43.104278 |
|
MtAiiM |
BsYtnP+MtAiiM |
0.000 |
0.076523 |
5.050485 |
AsAhlD |
BsAiiA+BsYtnP |
0.000 |
0.076523 |
5.050485 |
|
MtAiiM |
Commercial |
9.000 |
0.100000 |
6.600000 |
AsAhlD |
BsAiiA+MtAiiM |
0.000 |
0.076523 |
5.050485 |
|
AsAhlD |
SsAhlS |
3.000 |
0.653095 |
43.104278 |
AsAhlD |
BsYtnP+MtAiiM |
0.000 |
0.076523 |
5.050485 |
|
AsAhlD |
BsAiiA+BsYtnP |
0.000 |
0.076523 |
5.050485 |
AsAhlD |
Commercial |
9.000 |
0.100000 |
6.600000 |
|
AsAhlD |
BsAiiA+MtAiiM |
0.000 |
0.076523 |
5.050485 |
SsAhlS |
BsAiiA+BsYtnP |
0.000 |
0.076523 |
5.050485 |
|
AsAhlD |
BsYtnP+MtAiiM |
0.000 |
0.076523 |
5.050485 |
SsAhlS |
BsAiiA+MtAiiM |
0.000 |
0.076523 |
5.050485 |
|
AsAhlD |
Commercial |
9.000 |
0.100000 |
6.600000 |
SsAhlS |
BsYtnP+MtAiiM |
0.000 |
0.076523 |
5.050485 |
|
BsAiiA+BsYtnP |
BsAiiA+MtAiiM |
9.000 |
0.072198 |
4.765081 |
SsAhlS |
Commercial |
9.000 |
0.100000 |
6.600000 |
|
BsAiiA+BsYtnP |
BsYtnP+MtAiiM |
9.000 |
0.072198 |
4.765081 |
BsAiiA+MtAiiM |
BsYtnP+MtAiiM |
6.000 |
0.653095 |
43.104278 |
|
BsAiiA+BsYtnP |
Commercial |
9.000 |
0.076523 |
5.050485 |
BsAiiA+MtAiiM |
Commercial |
9.000 |
0.076523 |
5.050485 |
|
BsYtnP+MtAiiM |
Commercial |
9.000 |
0.076523 |
5.050485 |
The U values range from 0.000 to 9.000, with many comparisons having U values close to 0 or 9, indicating either no difference or a large difference between groups, respectively. Some comparisons show large differences with low p-values (e.g., WB600 vs. BsYtnP, U = 0.000, p = 0.076523; BsAiiA vs. BsYtnP, U = 9.000, p = 0.076523). These may warrant further investigation. Other comparisons show no significant differences (e.g., WB600 vs. AtAttM, U = 1.000, p = 0.164160; AsAhlD vs. SsAhlS, U = 3.000, p = 0.653095).
In summary, we can conclude that the experimental groups (various transgenic B. subtilis strains and their combinations) differed significantly from the control group formulations (WB600 and pHT43) in their ability to degrade artificially added C4-HSL. Based on the p-values, it is evident that most transgenic B. subtilis strains can degrade C4-HSL (with relatively small p-values), with BsAiiA being the most effective.
4.2 Natural AHL degradation test
Normality
The Shapiro-Wilk W value of A. hydrophila is extremely close to 1, and the p-value is very huge, as seen in the table. It suggests that the distribution of the A. hydrophila is normal. Since the WB600 and the A. hydrophila are similar, it can be claimed that the WB600 also has a normal distribution. As the AtAttM likewise follows a normal distribution, the W value and p-value are similar to those of the A. hydrophila and WB600. The A. hydrophila, WB600, AtAttM, and commercial strain all strongly follow a normal distribution.
Table10: Shapiro-Wilk test
|
Group |
Shapiro-Wilk W Statistics |
P-value |
Normality |
|
A. hydrophila |
1.0 |
1.0 |
yes |
|
WB600 |
1.0 |
1.0 |
yes |
|
pHT43 |
0.923 |
0.463 |
yes |
|
AtAttM |
1.0 |
1.0 |
yes |
|
Commercial |
1.0 |
1.0 |
yes |
Homogeneity of Variance Test
Table11: Levene test
|
Group |
Levene Statistics |
p-value |
Homogeneity of variance judgment |
|
All groups |
0.9246231155778888 |
0.48718870089575417 |
satisfy |
The Levene statistic is 0.925 with a p-value of 0.487, indicating no significant difference in variances between the groups (p > = 0.05).
ANOVA Test
Table12: ANOVA Test
|
Group |
F Statistics |
P-value |
Results |
|
All groups |
0.7534246575342463 |
0.5781313628744673 |
There are not significant differences between the means of the data sets |
With a p-value of 0.578 and the F-test statistic of 0.753, the results surpass 0.05. This suggests that the four datasets (A. hydrophila, WB600, pHT43, AtAttM, and the commercial strain) do not significantly differ from one another.
4.3 Biofilm reduction test
Normality Analysis
Table 13: Shapiro-Wilk test
|
Group |
Shapiro-Wilk statistics |
p-value |
Normality |
|
A. hydrophila |
0.8809148182824176 |
0.31347709322099665 |
yes |
|
WB600 |
0.7396666424158596 |
0.02381042591082771 |
no |
|
pHT43 |
0.783096360233108 |
0.058586241364745904 |
yes |
|
BsAiiA |
0.9406357478616523 |
0.6704364978873675 |
yes |
|
BsYtnP |
0.9374149723300009 |
0.6477246626054861 |
yes |
|
MtAiiM |
0.9717827663737167 |
0.8865679632958554 |
yes |
|
AtAttM |
0.8734550234369988 |
0.2806970317452857 |
yes |
|
AsAhlD |
0.8917172431338627 |
0.3658119652679628 |
yes |
|
SsAhlS |
0.794838020304235 |
0.07350401496404631 |
yes |
|
BsAiiA+BsYtnP |
0.8885980546166782 |
0.35010277677438173 |
yes |
|
BsAiiA+MtAiiM |
0.9624034540090122 |
0.824613077185083 |
yes |
|
BsYtnP+MtAiiM |
0.9161923016087565 |
0.5056799879670189 |
yes |
|
Commercial |
0.9777935406343107 |
0.9224849484049759 |
yes |
The majority of the data have a normal distribution; 12 of the 13 sample groups have normalcy, as indicated by p-values ≥ 0.05. These comprise the commercial strain as well as A. hydrophila, pHT43, BsAiiA, BsYtnP, MtAiiM, AtAttM, AsAhlD, and SsAhlS in combination.
There is one group that does not have a normal distribution: WB600 has a non-normality p-value of < 0.05 (0.024). Given that pHT43 and WB600 are both non-genetically changed B. subtilis strains, we can utilize pHT43 as the control group in the ensuing ANOVA study while ignoring WB600.
Homogeneity of Variance Test
Table 14: Levene test
|
Group |
Levene statistic |
P-value |
variance homogeneity judgment |
|
All group |
1.9369996642150542 |
0.05758118 |
meet |
The Levene statistic is 0.021, indicating minimal variance differences among groups. With a p-value greater than 0.05, we conclude that the variances are statistically equal, satisfying the assumption of homogeneity of variance.
ANOVA Test
Table 15: ANOVA Test
|
Group |
F-statistics |
p-value |
Results |
|
All group |
39.76951378076134 |
2.4494347373303303e-20 |
There are significant differences between the means of the data sets |
The F-value of 39.77 indicates high between-group variability compared to within-group variability. The p-value of approximately 2.45e-20 (<< 0.05) leads us to reject the null hypothesis, concluding that there are statistically significant differences in group means. This implies that the experimental groups differ significantly from the control group in their ability to inhibit A. hydrophila biofilm formation. Next, we will use post-hoc tests (e.g., Bonferroni correction) to identify which specific groups exhibit significant differences.
4. Post-Hoc Multiple Comparisons
Table 16: Tukey HSD test
|
Group 1 |
Group 2 |
Mean difference |
P-value |
Lower bound |
Upper bound |
significant |
|
A.h. |
AsAhlD |
-0.1994 |
0.0000 |
-0.2652 |
-0.1335 |
True |
|
A.h. |
AtAttM |
-0.0399 |
0.6391 |
-0.1057 |
0.0260 |
False |
|
A.h. |
BsAiiA |
-0.2119 |
0.0000 |
-0.2778 |
-0.1461 |
True |
|
A.h. |
BsAiiA+BsYtnP |
-0.2021 |
0.0000 |
-0.2680 |
-0.1363 |
True |
|
A.h. |
BsAiiA+MtAiiM |
-0.2106 |
0.0000 |
-0.2764 |
-0.1448 |
True |
|
A.h. |
BsYtnP |
-0.1975 |
0.0000 |
-0.2633 |
-0.1317 |
True |
|
A.h. |
BsYtnP+MtAiiM |
-0.1993 |
0.0000 |
-0.2651 |
-0.1334 |
True |
|
A.h. |
MtAiiM |
-0.2043 |
0.0000 |
-0.2701 |
-0.1385 |
True |
|
A.h. |
SsAhlS |
-0.1965 |
0.0000 |
-0.2623 |
-0.1306 |
True |
|
A.h. |
pHT43 |
-0.0280 |
0.9440 |
-0.0938 |
0.0378 |
False |
|
A.h. |
Commercial |
-0.0570 |
0.1493 |
-0.1228 |
0.0089 |
False |
|
AsAhlD |
AtAttM |
0.1595 |
0.0 |
0.0937 |
0.2253 |
True |
|
AsAhlD |
BsAiiA |
-0.0126 |
0.9999 |
-0.0784 |
0.0532 |
False |
|
AsAhlD |
BsAiiA+BsYtnP |
-0.0028 |
1.0 |
-0.0686 |
0.063 |
False |
|
AsAhlD |
BsAiiA+MtAiiM |
-0.0113 |
1.0 |
-0.0771 |
0.0546 |
False |
|
AsAhlD |
BsYtnP |
0.0018 |
1.0 |
-0.064 |
0.0677 |
False |
|
AsAhlD |
BsYtnP+MtAiiM |
0.0001 |
1.0 |
-0.0657 |
0.0659 |
False |
|
AsAhlD |
MtAiiM |
-0.005 |
1.0 |
-0.0708 |
0.0609 |
False |
|
AsAhlD |
SsAhlS |
0.0029 |
1.0 |
-0.0629 |
0.0687 |
False |
|
AsAhlD |
pHT43 |
0.1714 |
0.0 |
0.1055 |
0.2372 |
True |
|
AsAhlD |
Commercial |
0.1424 |
0.0 |
0.0766 |
0.2082 |
True |
|
AtAttM |
BsAiiA |
-0.1721 |
0.0 |
-0.2379 |
-0.1063 |
True |
|
AtAttM |
BsAiiA+BsYtnP |
-0.1623 |
0.0 |
-0.2281 |
-0.0965 |
True |
|
AtAttM |
BsAiiA+MtAiiM |
-0.1708 |
0.0 |
-0.2366 |
-0.1049 |
True |
|
AtAttM |
BsYtnP |
-0.1577 |
0.0 |
-0.2235 |
-0.0918 |
True |
|
AtAttM |
BsYtnP+MtAiiM |
-0.1594 |
0.0 |
-0.2252 |
-0.0936 |
True |
|
AtAttM |
MtAiiM |
-0.1645 |
0.0 |
-0.2303 |
-0.0986 |
True |
|
AtAttM |
SsAhlS |
-0.1566 |
0.0 |
-0.2224 |
-0.0908 |
True |
|
AtAttM |
pHT43 |
0.0119 |
1.0 |
-0.054 |
0.0777 |
False |
|
AtAttM |
Commercial |
-0.0171 |
0.9989 |
-0.0829 |
0.0487 |
False |
|
BsAiiA |
BsAiiA+BsYtnP |
0.0098 |
1.0 |
-0.056 |
0.0756 |
False |
|
BsAiiA |
BsAiiA+MtAiiM |
0.0013 |
1.0 |
-0.0645 |
0.0671 |
False |
|
BsAiiA |
BsYtnP |
0.0144 |
0.9998 |
-0.0514 |
0.0802 |
False |
|
BsAiiA |
BsYtnP+MtAiiM |
0.0127 |
0.9999 |
-0.0531 |
0.0785 |
False |
|
BsAiiA |
MtAiiM |
0.0076 |
1.0 |
-0.0582 |
0.0734 |
False |
|
BsAiiA |
SsAhlS |
0.0155 |
0.9996 |
-0.0503 |
0.0813 |
False |
|
BsAiiA |
pHT43 |
0.1839 |
0.0 |
0.1181 |
0.2498 |
True |
|
BsAiiA |
Commercial |
0.155 |
0.0 |
0.0892 |
0.2208 |
True |
|
BsAiiA+BsYtnP |
BsAiiA+MtAiiM |
-0.0085 |
1.0 |
-0.0743 |
0.0573 |
False |
|
BsAiiA+BsYtnP |
BsYtnP |
0.0046 |
1.0 |
-0.0612 |
0.0704 |
False |
|
BsAiiA+BsYtnP |
BsYtnP+MtAiiM |
0.0029 |
1.0 |
-0.0629 |
0.0687 |
False |
|
BsAiiA+BsYtnP |
MtAiiM |
-0.0022 |
1.0 |
-0.068 |
0.0636 |
False |
|
BsAiiA+BsYtnP |
SsAhlS |
0.0057 |
1.0 |
-0.0601 |
0.0715 |
False |
|
BsAiiA+BsYtnP |
pHT43 |
0.1741 |
0.0 |
0.1083 |
0.24 |
True |
|
BsAiiA+BsYtnP |
Commercial |
0.1452 |
0.0 |
0.0794 |
0.211 |
True |
|
BsAiiA+MtAiiM |
BsYtnP |
0.0131 |
0.9999 |
-0.0527 |
0.0789 |
False |
|
BsAiiA+MtAiiM |
BsYtnP+MtAiiM |
0.0114 |
1.0 |
-0.0545 |
0.0772 |
False |
|
BsAiiA+MtAiiM |
MtAiiM |
0.0063 |
1.0 |
-0.0595 |
0.0721 |
False |
|
BsAiiA+MtAiiM |
SsAhlS |
0.0142 |
0.9998 |
-0.0517 |
0.08 |
False |
|
BsAiiA+MtAiiM |
pHT43 |
0.1826 |
0.0 |
0.1168 |
0.2484 |
True |
|
BsAiiA+MtAiiM |
Commercial |
0.1537 |
0.0 |
0.0878 |
0.2195 |
True |
Table 16 reveals that most comparisons between the A. hydrophila group and other groups are statistically significant, indicating that all experimental groups inhibit A. hydrophila biofilm formation to some degree.
Our key findings include a) Significant differences are observed between AsAhlD, pHT43, the commercial strain, and various combinations of BsAiiA, BsYtnP, and MtAiiM. b) Comparing A.h. and BsAiiA, the mean difference is -0.2119 (95% CI: -0.2778 to -0.1461), indicating a significant inhibitory effect by BsAiiA (p ≤ 0.05). c) BsAiiA has a better inhibitory effect than AtAttM, with a mean difference of -0.1721. d) Comparisons with the pHT43 group show that AsAhlD and BsAiiA have the most significant differences, with mean differences of 0.1714 and 0.1839, respectively (both p-adj = 0.0).
In summary, with pHT43 as the control group, BsAiiA is the most effective group, showing the largest mean difference (0.1839) and statistical significance (p-adj = 0.0).
4.4 Extracellular protease reduction test
Normality Analysis
Table 17: Shapiro-Wilk test
|
|
Shapiro-Wilk staistics |
p-value |
Normality |
|
WB600 |
0.7960122247008959 |
0.10501335439912574 |
yes |
|
pHT43 |
0.9504797806716934 |
0.5714080891875228 |
yes |
|
BsAiiA |
0.9658288674879865 |
0.6449119636695078 |
yes |
|
BsYtnP |
0.8146817248459959 |
0.15005291013396682 |
yes |
|
MtAiiM |
0.9604128812773032 |
0.6174517146780358 |
yes |
|
AtAttM |
0.8492106384196234 |
0.23834132754163906 |
yes |
|
AsAhlD |
0.9974115616911132 |
0.902790698489526 |
yes |
|
SsAhlS |
0.9991917689752439 |
0.9456965208203199 |
yes |
|
BsAiiA+BsYtnP |
0.8356541071846479 |
0.2028018979666376 |
yes |
|
BsAiiA+MtAiiM |
0.9720704524976757 |
0.6793169812592069 |
yes |
|
BsYtnP+MtAiiM |
0.9733524735515158 |
0.686831718857623 |
yes |
Based on the p-values, all tests failed to reject the null hypothesis of normality. Therefore, all of our data is normal.
Homogeneity of Variance Test
Table 18: Levene test
|
Group |
Levene statistic |
P-value |
variance homogeneity judgment |
|
All groups |
8.78649008379585 |
0.04138153813433402 |
cannot be met |
The Levene statistic is 8.78649008379585, suggesting significant variations in variances between the groups. The p-value is less than 0.05, at 0.04138153813433402. we can draw the conclusion that the data do not meet the homogeneity of variance assumption and that there are notable variance differences between the groups.
ANOVA Test
Table 19: Welch's ANOVA TEST
|
Group |
F-statistics |
p-value |
Results |
|
All groups |
17.495783622510345 |
2.994090025819375e-08 |
There are significant differences between the means of the data sets |
When compared to within-group variability, the F-value of 17.496 shows that there is more between-group variability. The null hypothesis is rejected due to the p-value of roughly 2.99e-08 (<< 0.05). This suggests that the capacity of the experimental groups to lower A. hydrophila protease activity is considerably different from that of the control group. Post-hoc testing will then be used to determine which particular groups show significant differences.
Post-Hoc Multiple Comparisons
We will use the Games-Howell test to perform pairwise comparisons between each pair of sample groups. This is another multiple comparison method suitable for cases where variances are not homogeneous. It is similar to the Tukey HSD test but does not rely on the assumption of equal variances.
Table 20: Games-Howell test
|
Group 1 |
Group 2 |
Mean difference |
P-value |
Lower bound |
Upper bound |
significant |
|
AsAhlD |
AtAttM |
11.9833 |
0.0009 |
3.99 |
19.9767 |
True |
|
AsAhlD |
BsAiiA |
-3.52 |
0.8769 |
-11.5134 |
4.4734 |
False |
|
AsAhlD |
BsAiiA+BsYtnP |
-2.5267 |
0.984 |
-10.52 |
5.4667 |
False |
|
AsAhlD |
BsAiiA+MtAiiM |
-1.9833 |
0.9974 |
-9.9767 |
6.01 |
False |
|
AsAhlD |
BsYtnP |
-1.5733 |
0.9996 |
-9.5667 |
6.42 |
False |
|
AsAhlD |
BsYtnP+MtAiiM |
-1.7133 |
0.9992 |
-9.7067 |
6.28 |
False |
|
AsAhlD |
MtAiiM |
-1.6733 |
0.9994 |
-9.6667 |
6.32 |
False |
|
AsAhlD |
SsAhlS |
-0.6167 |
1.0 |
-8.61 |
7.3767 |
False |
|
AsAhlD |
WB600 |
11.1867 |
0.002 |
3.1933 |
19.18 |
True |
|
AsAhlD |
pHT43 |
13.65 |
0.0002 |
5.6566 |
21.6434 |
True |
|
AtAttM |
BsAiiA |
-15.5033 |
0.0 |
-23.4967 |
-7.51 |
True |
|
AtAttM |
BsAiiA+BsYtnP |
-14.51 |
0.0001 |
-22.5034 |
-6.5166 |
True |
|
AtAttM |
BsAiiA+MtAiiM |
-13.9667 |
0.0001 |
-21.96 |
-5.9733 |
True |
|
AtAttM |
BsYtnP |
-13.5567 |
0.0002 |
-21.55 |
-5.5633 |
True |
|
AtAttM |
BsYtnP+MtAiiM |
-13.6967 |
0.0002 |
-21.69 |
-5.7033 |
True |
|
AtAttM |
MtAiiM |
-13.6567 |
0.0002 |
-21.65 |
-5.6633 |
True |
|
AtAttM |
SsAhlS |
-12.6 |
0.0005 |
-20.5934 |
-4.6066 |
True |
|
AtAttM |
WB600 |
-0.7967 |
1.0 |
-8.79 |
7.1967 |
False |
|
AtAttM |
pHT43 |
1.6667 |
0.9994 |
-6.3267 |
9.66 |
False |
|
BsAiiA |
BsAiiA+BsYtnP |
0.9933 |
1.0 |
-7.0 |
8.9867 |
False |
|
BsAiiA |
BsAiiA+MtAiiM |
1.5367 |
0.9997 |
-6.4567 |
9.53 |
False |
|
BsAiiA |
BsYtnP |
1.9467 |
0.9978 |
-6.0467 |
9.94 |
False |
|
BsAiiA |
BsYtnP+MtAiiM |
1.8067 |
0.9988 |
-6.1867 |
9.8 |
False |
|
BsAiiA |
MtAiiM |
1.8467 |
0.9986 |
-6.1467 |
9.84 |
False |
|
BsAiiA |
SsAhlS |
2.9033 |
0.9595 |
-5.09 |
10.8967 |
False |
|
BsAiiA |
WB600 |
14.7067 |
0.0001 |
6.7133 |
22.7 |
True |
|
BsAiiA |
pHT43 |
17.17 |
0.0 |
9.1766 |
25.1634 |
True |
|
BsAiiA+BsYtnP |
BsAiiA+MtAiiM |
0.5433 |
1.0 |
-7.45 |
8.5367 |
False |
|
BsAiiA+BsYtnP |
BsYtnP |
0.9533 |
1.0 |
-7.04 |
8.9467 |
False |
|
BsAiiA+BsYtnP |
BsYtnP+MtAiiM |
0.8133 |
1.0 |
-7.18 |
8.8067 |
False |
|
BsAiiA+BsYtnP |
MtAiiM |
0.8533 |
1.0 |
-7.14 |
8.8467 |
False |
|
BsAiiA+BsYtnP |
SsAhlS |
1.91 |
0.9981 |
-6.0834 |
9.9034 |
False |
|
BsAiiA+BsYtnP |
WB600 |
13.7133 |
0.0002 |
5.72 |
21.7067 |
True |
|
BsAiiA+BsYtnP |
pHT43 |
16.1767 |
0.0 |
8.1833 |
24.17 |
True |
|
BsAiiA+MtAiiM |
BsYtnP |
0.41 |
1.0 |
-7.5834 |
8.4034 |
False |
|
BsAiiA+MtAiiM |
BsYtnP+MtAiiM |
0.27 |
1.0 |
-7.7234 |
8.2634 |
False |
|
BsAiiA+MtAiiM |
MtAiiM |
0.31 |
1.0 |
-7.6834 |
8.3034 |
False |
|
BsAiiA+MtAiiM |
SsAhlS |
1.3667 |
0.9999 |
-6.6267 |
9.36 |
False |
|
BsAiiA+MtAiiM |
WB600 |
13.17 |
0.0003 |
5.1766 |
21.1634 |
True |
|
BsAiiA+MtAiiM |
pHT43 |
15.6333 |
0.0 |
7.64 |
23.6267 |
True |
|
BsYtnP |
BsYtnP+MtAiiM |
-0.14 |
1.0 |
-8.1334 |
7.8534 |
False |
|
BsYtnP |
MtAiiM |
-0.1 |
1.0 |
-8.0934 |
7.8934 |
False |
|
BsYtnP |
SsAhlS |
0.9567 |
1.0 |
-7.0367 |
8.95 |
False |
|
BsYtnP |
WB600 |
12.76 |
0.0004 |
4.7666 |
20.7534 |
True |
|
BsYtnP |
pHT43 |
15.2233 |
0.0 |
7.23 |
23.2167 |
True |
|
BsYtnP+MtAiiM |
MtAiiM |
0.04 |
1.0 |
-7.9534 |
8.0334 |
False |
|
BsYtnP+MtAiiM |
SsAhlS |
1.0967 |
1.0 |
-6.8967 |
9.09 |
False |
|
BsYtnP+MtAiiM |
WB600 |
12.9 |
0.0004 |
4.9066 |
20.8934 |
True |
|
BsYtnP+MtAiiM |
pHT43 |
15.3633 |
0.0 |
7.37 |
23.3567 |
True |
|
MtAiiM |
SsAhlS |
1.0567 |
1.0 |
-6.9367 |
9.05 |
False |
|
MtAiiM |
WB600 |
12.86 |
0.0004 |
4.8666 |
20.8534 |
True |
|
MtAiiM |
pHT43 |
15.3233 |
0.0 |
7.33 |
23.3167 |
True |
|
SsAhlS |
WB600 |
11.8033 |
0.0011 |
3.81 |
19.7967 |
True |
|
SsAhlS |
pHT43 |
14.2667 |
0.0001 |
6.2733 |
22.26 |
True |
|
WB600 |
pHT43 |
2.4633 |
0.9867 |
-5.53 |
10.4567 |
False |
For pairwise comparisons, we employed the Games-Howell test, which works well in situations where the variances are not equal. There is a significant difference (p-adj = 0.002) between WB600 and AsAhlD, with WB600 having a higher mean. There is no difference between WB600 and AtAttM (p-adj = 1.0), with WB600 being somewhat lower. With the exception of WB600 (p-adj = 0.9867), pHT43 has a significantly higher mean (p-adj < 0.05) than almost all other groups.
These findings show that the remaining groups outperform WB600 in decreasing A.h. protease activity, with the exception of AtAttM and AsAhlD. Furthermore, when compared to the pHT43, all experimental groups demonstrate a greater and more notable capacity to decrease A. hydrophila protease activity.
5. Model 1 Conclusion
Through hypothesis testing, we found that our engineered B. subtilis significantly outperformed the unmodified control group in several key functional tests:
Synthetic and natural AHL degradation tests: The experimental groups, particularly the BsAiiA strain, significantly degrade the C4-HSL signal molecule.
Biofilm reduction test: The experimental group significantly reduces A. hydrophila biofilm formation, with strains like BsAiiA showing outstanding effects.
Extracellular protease reduction test: The experimental group significantly reduces A. hydrophila protease activity, mitigating its pathogenicity towards aquatic animals.
Model 2: Survival Analysis
In the previous chapter, we show the effectiveness of our method with statistics testing and reveal some characteristics (not directly inhibiting bacteria but preventing virulence element production). We aim to demonstrate its impact on mortality in aquaculture. So, in this chapter, we conduct a survival analysis.
Unfortunately, we are not able to directly collect mortality data due to a policy that prohibits animal experiments. Consequently, we decided to perform a survival analysis by referring to mortality data from similar released research studies (Chen et al., 2020; Chu et al., 2014).
Prior to utilizing the information from both referenced papers, we need to ensure that the information from both documents is highly relevant to our experiments.
a. Strain similarity: The strains used in these two papers (Bacillus sp. QSI-1 and B. licheniformis T-1) and our Wb800 both are closely related. For that reason, our stress must have similar features and potential antimicrobial mechanisms with them.
b. Similarities in experimental style: The two papers employed similar experimental designs, in which pathogenic bacteria (A. hydrophila) and potential probiotics (or stress with population quenching task) were infused right into zebrafish, and the survival case of experimental animals was observed.
1. Survival Analysis Curves
We plot survival curves, as shown in the following figure:
According to Figure 1, it is revealed that the probability of survival rate decreased over time in both the control group (blue line) and the QSI-1 group (orange line). This illustrates the threat of bacterial infection for aquaculture organisms (zebrafish). Furthermore, the survival probability of the QSI-1 group (orange line) is close to 1, indicating that the zebrafish in this group have a high survival rate. In the control group (blue line), the survival rate decreased considerably over time.
It is tentatively concluded that the strain with population quenching activity (QSI-1) provides protection against pathogen infection.
2. Log-Rank Test
To determine whether the difference in survival probability between the QSI-1 group and the control group is statistically significant, we conducted a log-rank test.
The log-rank test statistics are calculated as follows:
Statistic
where:
: the observed number of events in the QSI- 1 group
: the observed number of events in the Control group
: the expected number of events in the QSI-1 group
: the expected number of events in the Control group
Based on Table 21, the P-value (0.020921) is less than the significance level (0.05), we can conclude that the QSI-1 group including a strain with quorum quenching has a significant safety effect against microorganism infection compared to the control group. In summary, the above outcomes reveal a substantial distinction in zebrafish survival rates between both groups.
Table 21: log-rank test
|
test_statistic |
p-value |
-log2(p) |
|
5.333333 |
0.020921 |
5.578881 |
Model 2 Conclusion
In a word, the above results showed a significant difference in zebrafish survival rates between the two groups, indicating that the strain with quorum quenching technology protects against pathogen infection. Thus, developing engineered B. subtilis based on quorum quenching enzymes is feasible.
Conclusion
According to our experiment, the unmodified strain of B. subtilis WB600 performed worse in most cases compared to the engineered strains, particularly the one expressing the AHL lactonase gene BsAiiA. The experimental group significantly reduced the levels of the quorum-sensing signal molecule C4-HSL, which inhibited A. hydrophila from producing virulence factors, thereby reducing its pathogenicity. Additionally, the modified strains lowered protease activity and effectively inhibited the formation of biofilms. Our ANOVA and post-hoc analysis confirmed these findings. Furthermore, survival analysis provided insights into the long-term effects on animal survival rates. In conclusion, the engineered B. subtilis offers a promising alternative to traditional antibacterial methods.
References
Chen, B., Peng, M., Tong, W., Zhang, Q., & Song, Z. (2020). The Quorum Quenching Bacterium Bacillus licheniformis T-1 Protects Zebrafish against Aeromonas hydrophila Infection. Probiotics Antimicrob Proteins, 12(1), 160-171. https://doi.org/10.1007/s12602-018-9495-7
Chu, W., Zhou, S., Zhu, W., & Zhuang, X. (2014). Quorum quenching bacteria Bacillus sp. QSI-1 protect zebrafish (Danio rerio) from Aeromonas hydrophila infection. Sci Rep, 4, 5446. https://doi.org/10.1038/srep05446