I recently received the Organo-MS and ICP-MS results from five reef tanks under my care. Although I don't think many firm conclusions can be drawn from only five samples, I still find it quite interesting to compare these data with some more detailed information about the tanks.
Here's some information about the tanks:
(1) All of the tanks were a little over 2 years old, although most of the rock and substrate came from older tanks.
(2) No filter media other than the rock and substrate in the display had been used during the three months before sampling.
(3) Three tanks had only a skimmer in the sump, while the other two did not even have a skimmer, just an empty sump with water flowing through it.
(4) No amino acids were dosed in any of these tanks.
(5) There was no deliberate coral feeding. The fish were fed pellets and frozen mysids.
(6) None of the tanks had detectable nitrate or nitrite.
1. tank1
Skimmer: Yes
Dosing: All-For-Reef, sodium monobasic phosphate, ammonium bicarbonate
Water change: 20–30% every 1–3 months
Substrate: 1 inch of crushed coral
The tank is a mixed reef dominated mostly by stony corals, with a lot of Stylophora pistillata. Alkalinity consumption was about 1.5 dKH per day. The tank had been treated with fluconazole and oxolinic acid 7 months before sampling, and ciprofloxacin 3 months before sampling, to deal with a hair algae bloom and a coral disease outbreak.
2. tank2
Skimmer: Yes
Dosing: DIY two-part, Tropic Marin A- & K+, sodium monobasic phosphate, ammonium bicarbonate
Water change: 20–30% every 1–3 months
Substrate: 1 inch of crushed coral
The tank is an SPS-dominated reef. Alkalinity consumption was about 4 dKH per day. The tank had been treated with milbemycin oxime 7 months before sampling to treat parasitic copepods on Acropora spp.
3. tank3
Skimmer: No
Dosing: Kalkwasser, magnesium chloride
Water change: No water changes for at least a year
Substrate: 1 inch of crushed coral
The tank is a soft coral-dominated reef. Alkalinity consumption was about 0.5 dKH per day.
4. tank5
Skimmer: Yes
Dosing: DIY two-part, Tropic Marin A- & K+, sodium monobasic phosphate, ammonium bicarbonate, sodium nitrate (occasional boost)
Water change: 20–30% every 1–3 months
Substrate: Bare bottom
The tank is an SPS-dominated reef. Alkalinity consumption was about 4 dKH per day.
5. 4F tank
Skimmer: No
Dosing: All-For-Reef, sodium monobasic phosphate, ammonium bicarbonate
Water change: No water changes for at least a year
Substrate: Bare bottom
The tank is a mixed reef, but most of the coral biomass is probably contributed by the large Sarcophyton. The tank has experienced occasional bacterial blooms for no apparent reason ever since it was set up. Alkalinity consumption was about 0.1 dKH per day.
Amino acids
Most of the tanks showed a fairly similar overall pattern, characterized by high glycine and aspartic acid concentrations (Figure 1). I then calculated the total amino acid nitrogen concentration in these tanks, with 4F and tank3 having the highest concentrations (Figure 2). The proportion of total amino acid nitrogen contributed by each amino acid was then calculated (Figure 3). Interestingly, in the 4F tank, which had the highest total amino acid nitrogen concentration, the proportions of nitrogen contributed by aspartic acid and glycine were lower than in the other tanks.
Figure 1. Concentrations of different amino acids in five reef tanks
Figure 2. Total amino acid nitrogen in five reef tanks
Figure 3. Proportion of total amino acid nitrogen contributed by different amino acids in five reef tanks
Principal component analysis (PCA) of amino acid data
PCA is a dimensionality-reduction method that transforms a high-dimensional dataset, such as the amino acid concentrations and proportions here, into a lower-dimensional representation, in this case, two dimensions, for easier visualization. It is somewhat similar to projecting the shadow of your hand (a 3D object) onto a flat surface under a light source. The 2D shadow still preserves some useful characteristics of the original object. PCA basically finds the directions that capture the greatest amount of variation in the original dataset and represents the samples along these new axes. Here, the data are reduced to two axes, PC1 and PC2, which capture most of the variation among the samples.
I performed PCA on the amino acid mole fraction and molar concentration data (Figures 4 and 5). In both plots, 4F is farther away from the other tanks along the PC1 axis, which explains most of the variation in both analyses (63.44% and 70.12%). Figure 6 also shows the vectors for each amino acid in the mole-fraction PCA. The separation of 4F from the other tanks appears to be mostly associated with higher proportions of arginine, leucine + isoleucine, phenylalanine, tyrosine, and valine, while tank3 seems to be characterized by higher proportions of glycine and serine.
Figure 4. PCA plot of amino acid mole fractions in five reef tanks
Figure 5. PCA plot of amino acid concentrations in five reef tanks
Figure 6. PCA plot of amino acid mole fractions with amino acid vectors in five reef tanks
Vitamins & Plastic Additives
Figure 7 shows the concentrations of eight vitamins. Nicotinamide was detectable in all tanks, while riboflavin, cyanocobalamin, and thiamine were not detected in any of the five tanks. Tank1 had the most diverse vitamin composition of the five, including folic acid, pyridoxine, nicotinamide, and pantothenic acid.
As for plastic additives (Figure 8), only four were detected in these tanks: bisisobutylphthalate, bismethylphthalate, dibutylphthalate, and diethylphthalate.
Figure 7. Concentrations of various vitamins in five reef tanks
Figure 8. Concentrations of four plastic additives detected in five reef tanks
Plant-Derived Compounds
Four plant-derived compounds were detected at quantifiable concentrations in these tanks: coumarin, betaine, salicylic acid, and caffeine.
Figure 9. Concentrations of four plant-derived compounds detected in five reef tanks
Correlation Analysis Between Selected Parameters
Figure 10. Correlation between SAC254 and total amino acid nitrogen (p = 0.5860)
Figure 11. Correlation between SAC254 and glycine (p = 0.0227)
Figure 12. Correlation between phosphate and total amino acid nitrogen (p = 0.9450)
Figure 13. Correlation between DMSP and total amino acid nitrogen (p = 0.226)
Medication
Figure 14. Fluconazole and ciprofloxacin concentrations
Non-Targeted Screening
Here are a few things I found interesting:
1. Total amino acid nitrogen does not show much correlation with SAC254. However, if 4F is excluded, the relationship becomes a strong positive correlation, with a p-value of 0.02895.
2. Glycine is the only amino acid that shows a significant positive correlation with SAC254 without excluding any of the tanks.
3. Even though no inorganic nitrogen was detected in any of my tanks, total amino acid nitrogen still does not show any relationship with phosphate. I had originally expected at least a slightly negative correlation.
4. Aspartic acid and glycine concentrations in my tanks seem to be on the high end compared with reports from other reefers.
5. PCA of the amino acid mole fractions shows that the three stony coral-dominated tanks with skimmers and occasional water changes seem to have more similar amino acid compositions.
6. The amino acid composition of 4F is more enriched in several carbon-rich amino acids, such as Leu, Ile, Phe, Tyr, and Val.
7. Fluconazole, oxolinic acid, and ciprofloxacin had only been used to treat one of the tanks several months before the test, but several other systems still showed detectable amounts. This likely indicates that these compounds can be carried between systems fairly easily, possibly via rocks or corals.
8. Although the relationship was not statistically significant, DMSP did show a slightly negative correlation with total amino acid nitrogen.
9. Caffeine was detected in all tanks through non-targeted screening, which I'm not that surprised about since I do often drink coffee around these tanks. :)
If anyone wants to investigate these data further, I've also attached an XLSX file containing most of the raw data I received from Oceamo. I'd appreciate any feedback.
Here's some information about the tanks:
(1) All of the tanks were a little over 2 years old, although most of the rock and substrate came from older tanks.
(2) No filter media other than the rock and substrate in the display had been used during the three months before sampling.
(3) Three tanks had only a skimmer in the sump, while the other two did not even have a skimmer, just an empty sump with water flowing through it.
(4) No amino acids were dosed in any of these tanks.
(5) There was no deliberate coral feeding. The fish were fed pellets and frozen mysids.
(6) None of the tanks had detectable nitrate or nitrite.
1. tank1
Skimmer: Yes
Dosing: All-For-Reef, sodium monobasic phosphate, ammonium bicarbonate
Water change: 20–30% every 1–3 months
Substrate: 1 inch of crushed coral
The tank is a mixed reef dominated mostly by stony corals, with a lot of Stylophora pistillata. Alkalinity consumption was about 1.5 dKH per day. The tank had been treated with fluconazole and oxolinic acid 7 months before sampling, and ciprofloxacin 3 months before sampling, to deal with a hair algae bloom and a coral disease outbreak.
2. tank2
Skimmer: Yes
Dosing: DIY two-part, Tropic Marin A- & K+, sodium monobasic phosphate, ammonium bicarbonate
Water change: 20–30% every 1–3 months
Substrate: 1 inch of crushed coral
The tank is an SPS-dominated reef. Alkalinity consumption was about 4 dKH per day. The tank had been treated with milbemycin oxime 7 months before sampling to treat parasitic copepods on Acropora spp.
3. tank3
Skimmer: No
Dosing: Kalkwasser, magnesium chloride
Water change: No water changes for at least a year
Substrate: 1 inch of crushed coral
The tank is a soft coral-dominated reef. Alkalinity consumption was about 0.5 dKH per day.
4. tank5
Skimmer: Yes
Dosing: DIY two-part, Tropic Marin A- & K+, sodium monobasic phosphate, ammonium bicarbonate, sodium nitrate (occasional boost)
Water change: 20–30% every 1–3 months
Substrate: Bare bottom
The tank is an SPS-dominated reef. Alkalinity consumption was about 4 dKH per day.
5. 4F tank
Skimmer: No
Dosing: All-For-Reef, sodium monobasic phosphate, ammonium bicarbonate
Water change: No water changes for at least a year
Substrate: Bare bottom
The tank is a mixed reef, but most of the coral biomass is probably contributed by the large Sarcophyton. The tank has experienced occasional bacterial blooms for no apparent reason ever since it was set up. Alkalinity consumption was about 0.1 dKH per day.
Amino acids
Most of the tanks showed a fairly similar overall pattern, characterized by high glycine and aspartic acid concentrations (Figure 1). I then calculated the total amino acid nitrogen concentration in these tanks, with 4F and tank3 having the highest concentrations (Figure 2). The proportion of total amino acid nitrogen contributed by each amino acid was then calculated (Figure 3). Interestingly, in the 4F tank, which had the highest total amino acid nitrogen concentration, the proportions of nitrogen contributed by aspartic acid and glycine were lower than in the other tanks.
Figure 1. Concentrations of different amino acids in five reef tanks
Figure 2. Total amino acid nitrogen in five reef tanks
Figure 3. Proportion of total amino acid nitrogen contributed by different amino acids in five reef tanks
Principal component analysis (PCA) of amino acid data
PCA is a dimensionality-reduction method that transforms a high-dimensional dataset, such as the amino acid concentrations and proportions here, into a lower-dimensional representation, in this case, two dimensions, for easier visualization. It is somewhat similar to projecting the shadow of your hand (a 3D object) onto a flat surface under a light source. The 2D shadow still preserves some useful characteristics of the original object. PCA basically finds the directions that capture the greatest amount of variation in the original dataset and represents the samples along these new axes. Here, the data are reduced to two axes, PC1 and PC2, which capture most of the variation among the samples.
I performed PCA on the amino acid mole fraction and molar concentration data (Figures 4 and 5). In both plots, 4F is farther away from the other tanks along the PC1 axis, which explains most of the variation in both analyses (63.44% and 70.12%). Figure 6 also shows the vectors for each amino acid in the mole-fraction PCA. The separation of 4F from the other tanks appears to be mostly associated with higher proportions of arginine, leucine + isoleucine, phenylalanine, tyrosine, and valine, while tank3 seems to be characterized by higher proportions of glycine and serine.
Figure 4. PCA plot of amino acid mole fractions in five reef tanks
Figure 5. PCA plot of amino acid concentrations in five reef tanks
Figure 6. PCA plot of amino acid mole fractions with amino acid vectors in five reef tanks
Vitamins & Plastic Additives
Figure 7 shows the concentrations of eight vitamins. Nicotinamide was detectable in all tanks, while riboflavin, cyanocobalamin, and thiamine were not detected in any of the five tanks. Tank1 had the most diverse vitamin composition of the five, including folic acid, pyridoxine, nicotinamide, and pantothenic acid.
As for plastic additives (Figure 8), only four were detected in these tanks: bisisobutylphthalate, bismethylphthalate, dibutylphthalate, and diethylphthalate.
Figure 7. Concentrations of various vitamins in five reef tanks
Figure 8. Concentrations of four plastic additives detected in five reef tanks
Plant-Derived Compounds
Four plant-derived compounds were detected at quantifiable concentrations in these tanks: coumarin, betaine, salicylic acid, and caffeine.
Figure 9. Concentrations of four plant-derived compounds detected in five reef tanks
Correlation Analysis Between Selected Parameters
Figure 10. Correlation between SAC254 and total amino acid nitrogen (p = 0.5860)
Figure 11. Correlation between SAC254 and glycine (p = 0.0227)
Figure 12. Correlation between phosphate and total amino acid nitrogen (p = 0.9450)
Figure 13. Correlation between DMSP and total amino acid nitrogen (p = 0.226)
Medication
Figure 14. Fluconazole and ciprofloxacin concentrations
Non-Targeted Screening
| tank1 | tank2 | tank3 | tank5 | 4F |
| Fluconazole | Fluconazole | Fluconazole | Fluconazole | Fluconazole |
| Oxolinic acid | Oxolinic acid | Oxolinic acid | Oxolinic acid | Oxolinic acid |
| Crotamiton | Nicotinamide | Sucrose octaacetate | Sucrose octaacetate | Sucrose octaacetate |
| Cotinine N-oxide | (S)-Nicotine | (S)-Nicotine | Benzyl methacrylate | Benzyl methacrylate |
| Caffeine | Caffeine | Caffeine | Caffeine | Caffeine |
| Paraxanthine | Melamine | Melamine | Paraxanthine | Trometamol |
| Nicotinuric acid | Nicotinuric acid | Oridonin | Crotamiton | |
| Ciprofloxacin | Theobromine | Ciprofloxacin | Ciprofloxacin | |
| Carbendazim | Perfluoro-1-butanesulfonamide | |||
| 4-Pyridoxic acid | Perfluoro-1-butanesulfonic acid (PFBS) | |||
| (-)-Cotinine | Zearalanol | |||
| Perfluorohexanoic acid |
Here are a few things I found interesting:
1. Total amino acid nitrogen does not show much correlation with SAC254. However, if 4F is excluded, the relationship becomes a strong positive correlation, with a p-value of 0.02895.
2. Glycine is the only amino acid that shows a significant positive correlation with SAC254 without excluding any of the tanks.
3. Even though no inorganic nitrogen was detected in any of my tanks, total amino acid nitrogen still does not show any relationship with phosphate. I had originally expected at least a slightly negative correlation.
4. Aspartic acid and glycine concentrations in my tanks seem to be on the high end compared with reports from other reefers.
5. PCA of the amino acid mole fractions shows that the three stony coral-dominated tanks with skimmers and occasional water changes seem to have more similar amino acid compositions.
6. The amino acid composition of 4F is more enriched in several carbon-rich amino acids, such as Leu, Ile, Phe, Tyr, and Val.
7. Fluconazole, oxolinic acid, and ciprofloxacin had only been used to treat one of the tanks several months before the test, but several other systems still showed detectable amounts. This likely indicates that these compounds can be carried between systems fairly easily, possibly via rocks or corals.
8. Although the relationship was not statistically significant, DMSP did show a slightly negative correlation with total amino acid nitrogen.
9. Caffeine was detected in all tanks through non-targeted screening, which I'm not that surprised about since I do often drink coffee around these tanks. :)
If anyone wants to investigate these data further, I've also attached an XLSX file containing most of the raw data I received from Oceamo. I'd appreciate any feedback.
