Science of the Total Environment 615 (2018) 681–690

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Application of enzyme-hydrolyzed cassava dregs as a carbon source

in aquaculture
Qian Shang a,b,c,d, Haifang Tang d, Yinghui Wang a,b,c,⁎, Kefu Yu a,b,c, Liwei Wang a,b,c, Ruijie Zhang a,b,c,
Shaopeng Wang a,b,c, Rui Xue a,b,c, Chaoshuai Wei a,b,c
a
Guangxi Laboratory on the Study of Coral Reefs in the South China Sea, Guangxi University, 100 East Daxue Road, Nanning 530004, China
b
Coral Reef Research Center of China, Guangxi University, 100 East Daxue Road, Nanning 530004, China
c
School of Marine Sciences, Guangxi University, 100 East Daxue Road, Nanning 530004, China.
d
School of Environment, Guangxi University, 100 East Daxue Road, Nanning 530004, China

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• High yield of reducing sugar was pro-

duced by enzyme hydrolysis of cassava
dregs.
• Enzyme-hydrolyzed cassava dregs as a
substitute of sucrose applied in biofloc
aquaculture system.
• Enzyme-hydrolyzed cassava dregs was
more easily assimilation by heterotro-
phic bacteria.
• Microorganism in biofloc will affect the
indigenous microflora in shrimp
intestines.

a r t i c l e i n f o a b s t r a c t

Article history: As a kind of tropical agricultural solid waste, cassava dregs had become a thorny nonpoint source pollution prob-
Received 29 May 2017 lem. This study investigated the feasibility of applying cassava dregs as a substitute for sucrose in biofloc technol-
Received in revised form 23 August 2017 ogy (BFT) systems. Three types of biofloc systems (using three different carbon sources sucrose (BFT1), cassava
Accepted 26 August 2017
dregs (BFT2) and enzyme-hydrolyzed cassava dregs (BFT3) respectively), and the control were constructed in
Available online xxxx
this experiment in 200 L tanks with a C/N ratio of 20/1. The comparison of the water quality indicators (The
Keywords:
total ammonia nitrogen (TAN), nitrite (NO− −
2 -N), nitrate (NO3 -N), chemical oxygen demand (COD)), biofloc for

Enzyme-hydrolyzed cassava dregs the above four groups was performed, and the results indicated that BFT3 showed greater potential to the forma-
Carbon source tion of biofloc, which was beneficial for the water quality control. So the shrimp survival rate was the highest and
Biofloc the feed conversion rate was the lowest in BFT3. Besides, the high-throughput sequencing results showed that
Microbial communities the relative abundance of heterotrophic bacteria in the top 30 dominant microbial communities in BFT3 was
Probiotics higher than those in BFT1 and BFT2 by 20.70% and 1.19%, respectively, which could decrease TAN to improve
the water quality. Overall, the results had proved that the cassava dregs of enzymes hydrolysis could be used
as an ideal and cheap carbon source in BFT.
© 2017 Elsevier B.V. All rights reserved.

⁎ Corresponding author at: Guangxi Laboratory on the Study of Coral Reefs in the South China Sea, Guangxi University, 100 East Daxue Road, Nanning 530004, China.
E-mail address: wyh@gxu.edu.cn (Y. Wang).

https://doi.org/10.1016/j.scitotenv.2017.08.256
0048-9697/© 2017 Elsevier B.V. All rights reserved.
682 Q. Shang et al. / Science of the Total Environment 615 (2018) 681–690
1. Introduction
Litopenaeus vannamei (L. vannamei), commonly known as Pacific
white shrimp, is widely cultured in Asia, South America and North
America. The production of L. vannamei is responsible for approximately
76% of the total global shrimp yield in 2011 (Yu et al., 2014). Presently,
traditional L. vannamei aquaculture faces serious problems. The scarcity
and expense of water have limited aquaculture development. With the
expansion of the shrimp industry, the discharge of nutrient-rich efflu-
ents from culture ponds to coastal waters is becoming a major environ-
mental concern (Naylor et al., 1998; Shang et al., 1998). Moreover,
severe outbreaks of infectious diseases require more stringent
biosecurity measures, such as the reduction of water exchange rates
(Avnimelech and Kochba, 2009).
Currently, biofloc technology (BFT) is considered as an environmen-
tally friendly alternative for the development of intensive aquaculture
because it offers several advantages over conventional systems, such
as lower water renewal rates and a lower level of maintenance
(Haslun et al., 2012). Bioflocs are composed of algae, fungi, bacteria, fla-
gellates, ciliates and other microorganisms that adhere to an organic
matrix and present as floating flocs in the water column (Schryver
et al., 2008). In BFT technology, microorganisms maintain water quality
and serve as food for aquatic products. A BFT culture system contributes
to water quality by removing nitrogen compounds, supplementing the
diet of the aquatic animals and allowing for high stocking densities
(Krummenauer et al., 2016).
During the aquaculture period, shrimp excrete ammonia nitrogen,
which accumulates in ponds. A major source of ammonia nitrogen is
the typically protein-rich feed (Avnimelech, 1999). Mineralization of ac-
cumulated organic matter under anaerobic conditions leads to the for-
mation of toxic metabolites, such as NH3 and NO− 2 -N, and exposure to
toxic materials endangers the well-being of the cultured shrimp
(Avnimelech and Ritvo, 2003). However, nitrogen is used to produce
microcells that are rich in protein, and inorganic nitrogen is
immobilized into bacterial cells at a high C/N level. Controlling C/N
levels is a major focus of BFT research. Maintaining the C/N ratio within
a certain range is a prerequisite for floc formation (Avnimelech, 2006).
The C/N level of bacterial cells is 5:1 (Mooyoung et al., 2001), and the
carbon is used for respiration; thus, the optimal C/N level of the organic
substrates may be higher than that in the body composition under aer-
obic conditions. The C/N level in an aquaculture system could be in-
creased by adding different locally available inexpensive carbon
sources or by reducing the protein content in feed (Avnimelech, 1999;
Hargreaves, 2006).
At present, the mature carbon sources of BFT are sucrose, starch and
corn. However, the high cost of these carbon sources limits their use in
practice. Cassava dregs, a by-product of cassava starch (Manihot
esculenta Crantz) production, are thrown into water courses or left in
ditches, which overflow and carry much of the organic load towards
bodies of water. Their high water content and rich nutrients make cas-
sava dregs highly likely to cause microbial contamination of the sur-
rounding environment, leading to undesirable fermentation. Rainfall
create a serious non-point source pollution problem, which can pose a
threat to surrounding soil, rivers and human. (Fiorda et al., 2015) Sur-
prisingly, cassava dregs are an inexpensive carbohydrate a potential car-
bon source for BFT systems. Cassava dregs can be biologically converted
to various high-value products to maximize the effective utilization of
this important bioresource (Zhang et al., 2016b). N6 million tons of cas-
sava have been harvested for starch extraction since 2009 in Guangxi,
China (Sun et al., 2012), and approximately 700 kg of cassava dregs
with 70% water content are produced per 1000 kg cassava. These
dregs are difficult to degrade using natural environmental methods
alone; thus, it is extremely important to find more efficient and more
environmental friendly ways to use cassava dregs. The composition of
cassava dregs is complex, and starch, cellulose and lignin account for
61.3%, 10.5% and 8.6% respectively (Table A.1); however, utilizing all
of these contents is difficult. Studies has reported that cassava dregs
can be decomposed into small-molecule reducing sugars via
enzymolysis, the effects of which were distinct (Zhou et al., 2014). In
this study, we hydrolyzed cassava dregs using enzymolysis and then
added the innocuous and nonpoisonous products of this reaction to a
biofloc aquaculture system to replace the more expensive carbon
source, sucrose. The aim of this work was to explore the feasibility of
using cassava dregs as an alternative carbon source in biofloc
aquaculture.
An important step towards understanding and utilizing BFT was to
characterize and control the microbial community. In this study, we ex-
plore the changes in microbial aquaculture populations, investigate the
structure and diversity of the microbial communities in BFT and shrimp
intestine samples from L. vannamei culture ponds using high-
throughput sequencing technology. Our results clarified the relation-
ship among environmental factors and the microbial communities in a
biofloc aquaculture system as well as the microbial communities in
the intestine of L. vannamei. Our results provide novel information
that can be used to improve the microflora in aquaculture environments
and inhibit pathogens to reduce L. vannamei disease by adding new car-
bon sources to control the C/N ratio. These findings provide novel in-
sights into the feasibility of using cassava dregs as a carbon source in
biofloc aquaculture and reference the theory of BFT using the latest mi-
crobial identification technology and analysis methods.
2. Materials and methods
2.1. Materials
L. vannamei were purchased from the breeding base of the Beihai
Fishery Seed Co. Ltd., Guangxi, China. Healthy shrimp of similar size
were chosen. The initial body weight and body length of the shrimp
were 8.01 ± 0.14 g and 9.03 ± 0.20 cm, respectively. Before the formal
experiment, the shrimp were fed basic feed for 14 days to adapt to the
environment and feed. The diet consisted of commercial shrimp feed
with a crude protein level of 42%. The feeding levels accounted for 3%
of the shrimp body weight per day. To balance the C/N ratio and encour-
age the development of heterotrophic bacteria in the BFT, three carbon
sources were added to obtain a C:N of 20:1 (Avnimelech, 1999). In addi-
tion, the experimental groups were fed basic feed supplemented with su-
crose (BFT1), cassava dregs (BFT2) and cassava dregs treated with
enzymes (BFT3), moreover, the control was fed without an additional car-
bon source. The calculation of the C:N ratio considered the nitrogen con-
tent in the feed, the quantity of food distributed in the tanks and the rate
of nitrogen excretion by the shrimp, as mentioned in the description the
biofloc culture method (Hari et al., 2006). The amounts of sucrose, cassava
dregs and enzyme-hydrolyzed cassava dregs were 190%, 206% and 206%
of the shrimp feed, respectively (Wang et al., 2015).
Cassava dregs were provided by the Nanning Fushu Starch Factory,
Nanning, China. The water content of the fresh cassava dregs was 70%.
We sterilized the cassava dregs for 30 min at 121 °C in high-
temperature and high-pressure sterilization pots and dried the cassava
dregs for 2 h at 120 °C in a thermostatic drum wind drying oven. A
20–40 mesh screen was used to filter the cassava dregs after processing
in a crusher. After physical processing, the cassava dregs were hydro-
lyzed for 24 h at 60 °C by cellulase, which had an enzyme activity of
400 U/mg; for 12 h at 90 °C by α-amylase, which had an enzyme activity
of 4000 U/g; and for 6 h at 60 °C by a saccharifying enzyme with an en-
zyme activity of 100,000 U/mg (Table A.2). The reducing sugar contents
of the sucrose, cassava dregs, and enzyme-hydrolyzed cassava dregs
were 100%, 14.7% and 83.8%, respectively.
2.2. Experimental setup
The aquaculture experiments were performed at the aquaculture
center of the School of Marine Sciences, Guangxi University. Twelve
 Q. Shang et al. / Science of the Total Environment 615 (2018) 681–690
polyester tanks (110 cm × 70 cm × 60 cm) with an effective water vol-
ume of 200 L were used to cultivate the shrimp. These tanks were aer-
ated continuously during the breeding process. Fifty shrimp were
placed in each tank, and the experiment was run in triplicate.
The study lasted 30 days. During the experiment, shrimp were fed
four times per day at 8:00–9:00, 12:00–13:00, 17:00–18:00 and
22:00–23:00.
During the entire rearing period, all tanks were aerated and agitated
continuously using air stones connected to an air pump, and the water
in the tanks was maintained at a temperature of 22–24 °C and pH of
8.1–8.3, with the tanks aerated continuously to maintain a dissolved ox-
ygen (DO) level above 6.0 mg L−1. No water exchange occurred during
the experiment; however, a proper amount of fresh water was supplied
to compensate for water loss because of evaporation and water quality
measurements.
2.3. Measurement indicators and methods
2.3.1. Water quality
Throughout the 30-day experiment, the water temperature, salinity,
DO and pH were measured daily at 8:00 using a YSI-6600 V2 multipa-
rameter water quality sonde (YSI Incorporated, Yellow Springs, OH,
USA). Water samples (200 mL) were collected every 3 days from each
tank at approximately 9:00 and filtered through 0.45 μm GF/C filter
paper under vacuum pressure. Half of the water samples were analyzed
spectrophotometrically for total ammonia nitrogen (TAN) and nitrite
nitrogen (NO− −
2 –N), and nitrate nitrogen (NO3 –N) chemical oxygen de-
mand (COD) according to the specification for marine monitoring Part
4: seawater analysis (GB17378.4–1998); and the remaining half was fil-
tered under vacuum pressure through pre-dried and pre-weighed
Whatman GF/C filter paper. The filter paper containing suspended ma-
terials was dried at 105 °C in an oven reaching a constant weight, and
the dried sample was weighed to 0.01 mg using a Mettler AC 100 bal-
ance (Azim and Little, 2008). The total suspended solids (TSS) was cal-
culated from the weight difference (Azim et al., 2008). The biofloc
volume (BFV) was determined on site using Imhoff cones weekly by
registering the volume taken in by the bioflocs in 1000 mL of tank
water after 30 min of sedimentation (Yoram and Malka, 2009).
2.3.2. DNA extraction of biofloc samples and shrimp intestinal samples
At the end of the experiment, the microorganisms were collected by
0.22 μm membrane, and the genomic DNA of biofloc was extracted with
a Tissue DNA Kit (50) (Omega Bio-tek, Inc.). The shrimp were not fed for
24 h before sampling. Three shrimp specimens were randomly sampled
from each tank, and their intestines were aseptically dissected and
placed directly into 1.5 mL micro centrifuge tubes. Genomic DNA was
extracted from the intestinal samples with a Fast DNA® SPIN Kit for
Soil (MP Bio), and the extracted DNA was treated by vacuum freeze dry-
ing. All extracted DNA samples from the biofloc and shrimp intestines
were sent to Novogene Bioinformatics Technology Co., Ltd. for
sequencing.
2.3.3. Evaluation of richness and diversity
Mothur software was used to analyze alpha diversity (Bowman
et al., 2012). The community diversity in the samples was estimated
by Shannon-Wiener indices (Shannon, 2001), and rarefaction curves
were used to assess the species richness. The total number of species
in the samples (an exponent of the operational taxonomic units (OTU)
number) was estimated with Chao1 (Chao, 1984) and Ace (Pitta et al.,
2010). The sampling coverage was used to evaluate the authenticity of
the sequenced data (Singh et al., 2010).
2.3.4. Canoco analysis
The relationship between the designated microbial communities in
the samples and water quality indicators was analyzed by multivariate
statistical analysis software (Canoco for Windows 4.5). The program
683
WcanoImp of the Canoco for Windows 4.5 software package (Braak
and Smilauer, 2002) was used to generate spe.dta and env.dta files,
and a detrended correspondence analysis (DCA) was performed. A ca-
nonical correspondence analysis (CCA) or redundancy analysis (RDA)
was used for ordination according to the gradient lengths (an RDA
was used when the gradient length was b 3, an RDA or a CCA was used
when the gradient lengths were between 3 and 4, and a CCA was used
when gradient length was N4) to address the relationship between mi-
crobial communities and water quality indicators.
2.4. Statistical analysis
Data obtained from the experiment were analyzed using SPSS 17.0
software (SPSS, Chicago, USA) for Windows. A one-way ANOVA was
performed on the experimental parameters. When possible, all data
were statistically analyzed to detect any significant differences at the
5% level. When significant differences were found, Duncan's multiple
range test was used to identify the differences between the experimen-
tal groups. Different carbon sources and water quality indicators were
compared using multiple analyses of variance (ANOVAs) (Steicke
et al., 2007).
3. Results
3.1. Water quality and bioflocs development
The measured water quality in all experimental groups remained
within recommended levels for shrimp culture throughout the 30-day
experimental period, and the water quality parameters (temperature,
salinity, DO and pH) was consistent. The fluctuations of TAN, NO− 2 -N,
NO− 3 -N, throughout the experimental period were shown in Fig. 1a, b,
c. The result demonstrated that all of the nitrogen compounds concen-
tration in treatments were lower than the control (the final concentra-
tion of TAN, NO− −
2 -N and NO3 -N in the control was 0.38 ± 0.03, 0.43 ±
−1
0.03 and 17.80 ± 3.32 mg L respectively). Moreover, BFT1 had a bet-
ter effect on the controlling of TAN and NO−2 -N concentrations (0.16 ±
0.01 and 0.20 ± 0.01 mg L−1 respectively) compared with the other BFT
groups, and a significant difference was observed in the treatment effect
of cassava dregs with or without enzyme hydrolysis. Fluctuations of
COD, TSS, and BFV were shown in Fig. 2a, b, c. All of the COD concentra-
tion in the treatments and the control were b14 mg L−1. The final values
of TSS and BFV in BFT2 (247.55 ± 16 mg L−1 and 16.53 ± 1.07 mL L−1)
was higher than that in BFT3 Despite differences in water quality pa-
rameters among treatments, all of them were within acceptable ranges
for most penaeid species.(Wickins, 1976; Van Wyk et al., 1999;
Emerenciano et al., 2012).
3.2. Growth performance and feed utilization of shrimp
Growth performance was evaluated through final weight gain
(expressed as a percent of initial body weight) and survival rate. The
growth of the shrimp in both bioflocs treatments was significantly bet-
ter than that obtained in the control (Table 1). The feed conversion rate
of the shrimp in both bioflocs treatments was significantly lower than
that obtained in the control (Table 1). Survival rate of shrimp (44.0%,
65.33%, 54.67% and 66.67% for the control, BFT1, BFT2 and BFT3 respec-
tively) and final length differ between treatments and the control.
3.3. Bacterial diversity and composition in biofloc samples
High-throughput sequencing yielded 53,501, 40,037, 34,099 and
30,409 high-quality and effective reads for samples from the control,
BFT1, BFT2 and BFT3 respectively. The number of OTUs identified at
97% similarity in the samples from the control, BFT1, BFT2 and BFT3
were 887, 810, 722 and 585, respectively. In all samples, the detected
684 Q. Shang et al. / Science of the Total Environment 615 (2018) 681–690
Fig. 1. Effects of different carbon sources addition on the concentration of the (a) TAN,
(b) NO− −
2 -N, (c) NO3 -N.
bacteria were classified into 31 phyla, 57 classes, 109 orders, 182 fami-
lies and 360 genera.
3.4. Analysis of species abundance at the phyla level
The number of bacterial phyla in the control, BFT1, BFT2 and BFT3
were 23, 21, 23 and 23, respectively. Although the bacterial diversity
was similar, the frequency distribution of bacterial phyla differed
among the treatments as shown in Fig. 3. Proteobacteria and
Bacteroidetes were the dominant bacteria in aquaculture and occupied
Fig. 2. Effects of different carbon addition on the concentration of the (a) TSS, (b) BFV and
(c) COD.
a very large proportion in all types of biofloc. The dominant bacteria in
BFT3 clearly differed from those in the control, BFT1 and BFT2, and the
subdominant bacterium of BFT3 was Firmicutes. The relative abundance
of Proteobacteria in all samples was the control b BFT1 b BFT2 b BFT3,
suggesting that the microbial community was shaped by the carbon
source. A similar result was reported in previous research (Martins
et al., 2013). The majority of Proteobacteria detected in all groups were
considered to be symbiotic bacteria in aquaculture (Sakami et al.,
2008). Proteobacteria were known to remove organic matter (Wagner
et al., 1993; Miura et al., 2007), particularly in the wastewater treatment
of biofloc, which played a major role in the composition of these
 Q. Shang et al. / Science of the Total Environment 615 (2018) 681–690
Table 1
Growth performance and feed utilization of Litopenaeus vannamei in the control and three biofloc treatments with different carbon sources at the end of 30-day feeding experiment.
Parameter Control Sucrose
Final individual weight (g) 10.35 ± 0.33 11.50 ± 0.21
Weight gain (%) 29.70 ± 2.10 42.64 ± 2.0
Final length (mm) 11.51 ± 0.45 11.39 ± 0.27
Feed conversion rate 2.91 ± 0.04 1.87 ± 0.08
Survival rate (%) 44.0 ± 4.0 65.33 ± 6.11
bacteria (Wagner et al., 1995; Daims et al., 1999). Such results demon-
strated that when a biofloc was applied to a culture system, it could ef-
fectively regulate the quality of aquaculture water. Most notably, this
phylum was widely dispersed in the marine environment and played
an important role in the process of nutrient cycling and the mineraliza-
tion of organic compounds (Kirchman, 2002; Wei et al., 2016).
The relative abundance of Bacteroidetes was the control N BFT1
N BFT2 N BFT3, and this phylum represents a dominant member of ma-
rine heterotrophic bacterioplankton and frequently colonizes macro-
scopic organic matter particles (Martins et al., 2013; Woebken et al.,
2007). Bacteroidetes was the subdominant microbial phylum in the con-
trol, BFT1 and BFT2 samples, and Firmicutes was the subdominant phy-
lum in the BFT3 samples, with relative abundances that ranged from
7.2% in BFT3 to approximately 1% in the other groups. Altogether,
these three phyla represented N 90% of the total bacteria. The relative
abundance of Saccharibacteria in BFT3 was higher than that in BFT2, al-
though these bacteria were not found in BFT1. Actinobacteria were well
known for the production of secondary metabolites, including many po-
tent antibiotics (Ventura et al., 2007). Actinobacteria was a common pro-
biotic that can produce beneficial substances in certain environments
(Das et al., 2008), and the difference in the relative abundance of
Actinobacteria in all groups was not significant.
3.5. Analysis of species abundance at the family level
Although the compositions of bacteria in the biofloc at the phylum
level among samples were quite similar, the differences in composition
at the family and genus levels contributed to significant differences
among the different treatments. The top 30 bacterial reads at the family
level were defined as the dominant microbial family. The species abun-
dance in different samples was shown in Fig. 4. At the family level, the
bacteria with the highest abundance were Piscirickettsiaceae (13.38%)
in the control, GR-WP33-58 (14.7%) in BFT1, GR-WP33-58 (20.2%) in
BFT2 and Halomonadaceae (21.0%) in BFT3. The next most common
were Flavobacteriaceae (12.74%), Rhodobacteraceae (6.0%),
Fig. 3. Relative abundance of the most frequently identified bacterial phyla (N0.1% of total
sequence) in water from different rearing condition.
685
Cassava dregs Enzyme-hydrolyzed cassava dregs
10.76 ± 0.15 11.70 ± 0.44
34.51 ± 2.30 46.07 ± 3.20
11.53 ± 0.16 11.56 ± 0.19
2.09 ± 0.03 1.75 ± 0.07
54.67 ± 6.11 66.67 ± 3.06
Halomonadaceae (6.3%) and Rhodobacteraceae (14.3%) in the control,
BFT1, BFT2 and BFT3, respectively.
In the control, Piscirickettsiaceae and Flavobacteriaceae both were po-
tential pathogenic bacteria to shrimp, suggesting that the water quality
had deteriorated to pose a threat to shrimp health. As the dominant bac-
terial community in BFT1 and BFT2, GR-WP33-58 belongs to the order
Desulfuromonadales. These clones were divergent and form a monophy-
letic cluster with the environmental sequence GR-WP33-58 that
branched at the base of myxobacteria. The closest match was the envi-
ronmental sequence GR-WP33-58, which was retrieved from a uranium
waste pile (Moreira et al., 2006). Myxobacteria were gram-negative pro-
karyotes in the delta group of Proteobacteria. Halomonadacea were aer-
obic heterotrophic microorganisms that exhibit an extreme salt
tolerance.
Rhodobacteraceae which had been found at a high frequency in bio-
film systems (Hargreaves, 2006; Wang et al., 2015), was considered an
excellent biofilm-forming organism and represents one of the first and
dominant colonizers of the surface in all marine environments (Azim
and Little, 2008; Clescerl, 1998). The ability to effectively colonize sur-
faces gave Rhodobacteraceae a competitive advantage in bioflocs, and
the TSS concentration was important. In addition, the Rhodobacteraceae
family was known to exhibit a diverse range of metabolic activity. Wang
et al. reported that the Rhodobacter group could establish an antagonis-
tic beneficial bacterial community in the rearing environment of turbot
larva and thereby limit the survival of pathogenic bacteria (Wang et al.,
2015). In the present study, Rhodobacteraceae might had therefore
played an important role in maintaining the health of the culture sys-
tem. The relative abundance of Rhodobacteraceae in the control, BFT1,
BFT2 and BFT3 was 6.13%, 6.03%, 5.02% and 14.35% respectively, it dem-
onstrated that the BFT3 had significant advantage in the limiting the
survival of pathogenic bacteria.
3.6. Effect of environmental factors on bacterial communities, probiotics
and pathogens
High variability in the most frequent bacterial communities was ev-
idenced over the course of rearing in BFT, which was likely because of
changes in biotic and abiotic environmental factors (Ramette, 2007).
To explore the effects of environmental factors on microbial communi-
ties, 30 dominant OTUs at the family level for each sample were selected
based on the results of the DCA (Table A.5). A species-environment cor-
relation represented the explained variance of each ordination axis to
the species, and the cumulative percentage variance of the species-
environment was used to denote the explained variance of the environ-
mental variables to species in the RDA. The ordination results based on
the RDA were shown in Table A6. The correlation coefficient for the first
and second ordination axes was 0 for both species and environmental
factors, indicating that the ordination results were credible (Zhang
et al., 2016a). Three samples were located in different areas of the ordi-
nation diagram (Fig. 5a), indicating that they had different bacterial
communities. The environmental factor that was most positively corre-
lated to the first ordination axis was TSS content, and the most negative-
ly correlated factor was NO− 3 –N. The species-environment correlation
coefficient for the first ordination axis was 1.0, and the cumulative per-
centage variance of the species-environment relationship was 61.5%
(Table A.6). The increased TSS level and decreased NO− 3 –N level
686 Q. Shang et al. / Science of the Total Environment 615 (2018) 681–690
Fig. 4. Relative abundance of the most frequently identified 30 bacterial family in biofloc samples from different rearing conditions.
increased bacterial richness in water. The second ordination axis was
positively correlated with TAN, NO− −
2 –N and NO3 –N levels and negative-
ly correlated with BFV and TSS. Based on the length of the connecting
wire, TSS had the greatest impact on the bacterial populations in
water followed by the TAN and NO− 2 –N levels and BFV. The result of
present study demonstrated that TSS played the most important role
in biofloc rearing, with the reduced water exchange, higher organic
matter inputs, and higher growth rates of heterotrophic bacteria con-
tributing to an increase in TSS (Schveitzer et al., 2013). Ebeling et al.,
(2006) reported that in BFT systems, the density of suspended solids
was the second most important limiting factor to increased productivity
after oxygen (Ebeling et al., 2006). The interaction between nutrients
and bacterial communities was sustained throughout the entire exper-
imental period. The high TAN concentration improved the survival of
bacterial communities; in addition, most of the heterotrophic bacteria
consumed TAN for growth.
The correlation analysis of 30 OTUs at the family level with environ-
mental factors (Fig. 5b) showed that different bacterial genera had dif-
ferent adaptabilities to water of varying quality. These dominant
bacterial genera in water were marked in the ordination diagram
(Fig. 5b), and values falling within a close range indicated similar corre-
lations for certain bacterial genera with the aquatic environment, such
as Halomonadaceae (2), Rhodobacteraceae (3), Vibrionaceae (4),
Shewanellaceae (5), Lachnospiraceae (8), SS1-B-06-26 (15), Enterobacte-
riaceae (24) and Ruminococcaceae (27). The abundances of these eight
types of bacteria were positively correlated with the COD, TSS and BFV.
According to the correlation analysis between the abundance of
pathogenic genera and different environmental factors, the abundance
of Vibrionaceae (4) was positively correlated with the COD, TSS and
BFV. Flavobacteriaceae (7) showed a similar correlation with the water
parameters as Rickettsiaceae (21) and Aeromonadaceae (22), which
were positively correlated with the TAN, NO− −
2 -N and NO3 -N levels
and negatively correlated with the COD, BFV and TSS. The abundance
of Piscirickettsiaceae (26) showed an inverse relationship. The abun-
dances of Rhodobacteraceae (3) and Enterobacteriaceae (24) was posi-
tively correlated with the TSS and BFV and negatively correlated with
NO− 3 -N. The abundances of Bdellovibrionaceae (13), Rhodospirillaceae
(18), and Flammeovirgaceae (19) showed a negative correlation with
the TAN, NO− −
2 -N, NO3 -N, TSS, COD and BFV. However, certain patho-
genic and probiotic bacteria located in a similar position in the ordina-
tion diagrams had a similar correlation with the environmental
variables (Table A.3). Heterotrophic bacteria, such as Halomonadaceae
(2), Lachnospiraceae (8), Ruminococcaceae (27), and Enterobacteriaceae
(24), were generally positively correlated with the BFV. These hetero-
trophic bacteria utilize TAN and a carbon source to produce energy for
reproduction, benefitting the growth of bioflocs.
3.7. Effect of the rearing environment on shrimp intestines microbiota
In the biofloc system, we focused on the bacterial distribution and
composition in shrimp intestines after 30 days of rearing under different
conditions.
High-throughput sequencing yielded 53,547, 59,512, 49,122, and
55,166 high-quality and effective reads for samples S0, S1, S2, and S3,
respectively. Samples S0, S1, S2 and S3 were intestinal samples from
the sucrose group, cassava dregs group and enzyme-hydrolyzed cassava
dregs group, respectively. The numbers of OTUs identified at 97% simi-
larity in samples S0, S1, S2 and S3 were 449, 471, 221 and 280, respec-
tively (Table A.4). In all samples, the detected bacteria were classified
into 24 phyla, 39 classes, 69 orders, 153 families and 220 genera. The
number of OTUs in the intestines and the diversity of the microbial com-
munity in the intestines were obviously less than that in the biofloc. The
number of gene reads in the sequenced intestinal samples were higher
than that in the biofloc samples. However, the variety of microbial com-
munities in all samples showed the opposite results, with the biofloc
consistently containing lower microbial biomass than the shrimp guts;
however, the 16S rDNA sequences indicated that the biofloc was more
diverse than the shrimp guts.
The Shannon index was 2.63, 3.27, 2.58, 3.48, 6.83, 7.57, 6.92 and
5.88 in S0, S1, S2, S3, the control, BFT1, BFT2 and BFT3, respectively,
and the Simpson index was 0.66, 0.83, 0.63, 0.84, 0.97, 0.98, 0.97 and
0.94 in S0, S1, S2, S3, the control, BFT1, BFT2 and BFT3, respectively.
These findings suggested a considerable difference in the microbial di-
versity in the intestines and biofloc. Bacterial diversity was estimated
by the Shannon index, where a higher value corresponds to high bacte-
rial diversity. The results revealed a significantly lower diversity in the
intestines compared with that in the biofloc. The Simpson index con-
firmed this result, with a higher value also indicating high bacterial
diversity.
The S0, S1, S2 and S3 libraried from intestine samples were mostly
dominated by one order belonging to Vibrionales, which represented
88.95%, 62.50%, 87.00% and 60.70% of the total sequences in the shrimp
intestines from S0, S1, S2 and S3, respectively. Fusobacteriales was the
second most abundant order, with a relative abundance of 0.04% for
 Q. Shang et al. / Science of the Total Environment 615 (2018) 681–690
Fig. 5. Correlation analysis between bacterial communities and environmental factors.
RDA (redundancy analysis) ordination diagrams of environmental factors and: a,
bacterial communities in biofloc samples; b, specific bacterial communities in biofloc
samples. 1 GR-WP33-58, 2 Halomonadaceae, 3 Rhodobacteraceae, 4 Vibrionaceae, 5
Shewanellaceae, 6 Nannocystaceae, 7 Flavobacteriaceae, 8 Lachnospiraceae, 9
Saprospiraceae, 10 1G93, 11 Bacteriovoracaceae, 12 Campylobacteraceae, 13
Bdellovibrionaceae, 14 Hyphomonadaceae, 15 SS1-B-06-26, 16 Alteromonadaceae, 17
unidentified_Gammaproteobacteria, 18 Rhodospirillaceae, 19 Flammeovirgaceae, 20
Cryomorphaceae, 21 Rickettsiaceae, 22 Aeromonadaceae, 23 unidentified_Order_III, 24
Enterobacteriaceae, 25 JTB255_marine_benthic_group, 26 Piscirickettsiaceae, 27
Ruminococcaceae, 28 Erythrobacteraceae, 29 Colwelliaceae, 30 Kordiimonadaceae.
S0, 15.4% for S1, 1.1% for S2 and 24.4% for S3; thus, obvious differences
occurred among the three groups. The relative abundance of
Alteromonadales was similar in the different samples and ranged from
1.9% to 2.1%. Rhodobacterales was the third most abundant order in S3
(7.90% of total sequences). Oceanospirillales showed concentrations of
0.27%, 1.40%, 3.83%, 1.22% in S0, S1, S2 and S3, respectively. (See Fig. 6.)
As shown in Fig. 7 the dominant bacterial genera in the intestinal
population of each shrimp varied among the different treatments,
these genera were Photobacterium, Vibrio, Shewanella, and
Propionigenium (phylum Fusobacteria). Similar results had been report-
ed for Peneaus monodon (Chaiyapechara et al., 2012). In the intestinal
samples, Vibrionaceae was represented by Photobacterium damselae.
This species was a known marine fish pathogen that could be transmit-
ted to fish through water, and this group of bacteria had previously been
reported at a high relative abundance in the intestines of other species
of shrimp, such as L. vannamei, P. monodon and Fenneropenaeus chinensis
(Chaiyapechara et al., 2012; Rungrassamee et al., 2013). Photobacterium
687
Fig. 6. Relative abundance of the most frequently identified bacterial orders within shrimp
intestines from two different types of rearing conditions.
represented 55.32%, 27.70%, 59.23% and 22.51% of Vibrionaceae in the
total sequence of all shrimp intestines. Clearly, the abundance in the
S1 and S3 samples was far lower than that in S0 and S2, and the quanti-
ties of Photobacterium in the S1 and S3 samples were effectively con-
trolled. The contents of Propionigenium in S0, S1, S2 and S3 were
0.04%, 15.40%, 1.12%, 24.01%, respectively. Other genera found in signif-
icant proportions in the intestinal bacterial community included
Ruegeria, Mycoplasmataceae, Fusibacter, Halomonas, and Candidatus
Bacilloplasma. These intestinal bacterial communities varied significant-
ly among different treatments and were distinct from their rearing
water.
4. Discussions
As shown in Fig. 1a, b, c, the indices of nutrients in the control were
higher than three treatments, so the pressure from adverse environ-
mental condition of the control was more serious than treatments,
and it also demonstrated that sucrose and enzyme-hydrolyzed cassava
dregs both had better effect on the controlling of nutrients. It had
been shown that there were different effects of simple versus complex
carbohydrates used as the carbon source in biofloc systems. Simple
sugars resulted in faster ammonia removal, while more complex carbo-
hydrates required more time for decomposition into simple sugars,
thereby resulting in slower ammonia removal. Such results might ex-
plain the higher TAN level in BFT2 compared with BFT1 and BFT3
(Kirchman, 2002). Each type of biofloc grown on different carbon
sources (sucrose, cassava dregs and enzyme-hydrolyzed cassava
dregs) was adequate in maintaining the water quality parameters in a
normal range for shrimp growth (Clescerl, 1998; Van Wyk et al., 1999).
As shown in Fig. 1a, the TAN increased gradually until the 9th day in
all treatments, then the increase became slowly, even decreased in BFT3
during the 9th to 15th day; While in Fig. 1b, there was a sharp rise of
NO− −
2 -N and NO3 -N from the 9th day and 12th day, respectively, such
results demonstrated that nitrification had taken on. Such result was
consistent with the conclusion of previous research that the immobili-
zation of ammonia nitrogen by nitrifying bacteria occurred much
more slowly, because the growth rate and microbial biomass per unit
substrate of heterotrophic bacteria were tenfold higher than those of ni-
trifying bacteria (Hopkins et al., 1994). During the 12th day and the 15th
day, the growth rate of NO− 3 -N in all groups became slowly in Fig. 1c,
even the control and BFT2 performed significant decrease, while both
the TAN and NO− 2 -N were descending, so it could be concluded that as-
similation by heterotrophic bacteria, nitrification by nitrifying bacteria
and denitrification by denitrifying bacteria were simultaneous in the
tanks. The obvious transition began from the 15th day, all of the
688 Q. Shang et al. / Science of the Total Environment 615 (2018) 681–690
Fig. 7. Relative abundance of 10 the most frequently identified bacterial genus within biofloc and bacterial genus within shrimp intestines from four different types of rearing conditions.
indicators increased almost at the same time in all groups, when the ef-
fect of the nitrification was better than denitrification. Another great
change of NO− −
2 -N and NO3 -N appeared during the 18th and 21th day,
and the NO− −
3 -N of all groups and NO2 -N of treatments decreased appar-
ently, while the TAN continued to increase, so it was obvious that the ef-
fect of the denitrification was better than nitrification. After the 21th
day the NO− 3 -N increased again, which demonstrated that the nitrifica-
tion became dominant. Moreover, the feeding amount of commercial
feed was reduced as the decrease of shrimp survival rate, thus the
TAN decreased rapidly in BFT1and BFT3 which had better biofloc.
The trend of COD in the control and treatments was consistent in the
aquaculture process, which was similar to previous research (Zhang
et al., 2016a, 2016b). TSS and BFV were often used as indicators for
quantitative determination of biofloc(Avnimelech, 1999), and it indicat-
ed that adding carbon source was benefit for the growth of biofloc in
Fig. 2. Both the BFV and TSS levels increased gradually throughout the
experiment, and the change tendency over time was basically consis-
tent. Apparently, BFT3 had a greater advantage than BFT1 and BFT2, be-
cause BFT3 had a high concentration of carbohydrates which could be
the results of the poor solubility of cassava dregs because cassava
dregs particles might still be present in the flocs. According to the re-
search of Gaona et al., the TSS and BFV values should not exceed
500 mg L−1 and 40 mg L−1, respectively (Gaona et al., 2011) and the re-
sults of this study was consistent with their conclusions.
As can be seen from Table 1, the growth indicators (in terms of final
weight, weight gain and survival rate) of the shrimp in all bioflocs treat-
ments were significantly higher than that obtained in the control, while
the feed conversion rate was significantly lower, and BFT3 exhibited the
best results among all of the treatments. Such results were in agreement
with the previous conclusions in L. vannamei with bioflocs present in
the culture systems (Wasielesky et al., 2006; Burford et al., 2004; Xu
and Pan, 2012). It was possible that the shrimp fed on both the commer-
cial feed and the bioflocs might had better nutrient assimilation when
compared to those fed on only the commercial feed, because of the
greater amount of essential amino acids, fatty acids (PUFA and HUFA)
and other nutritional elements supplied by the bioflocs (Izquierdo
et al., 2006; Ju et al., 2008; Agj et al., 2015). More importantly, the
bioflocs could exert a positive effect on digestive enzyme activity of
the shrimp, which made contribution to feed digestion and utilization
(Xu and Pan, 2012).
Previous studies showed that the structures of bacterial community
were closely to changes in the aquaculture environment (Payne et al.,
2007; Crystaln et al., 2010; Tomoko et al., 2010; Zeng et al., 2010), and
disease outbreaks were usually involved in the abundance of pathogen-
ic bacteria in aquacultures (Lavilla-Pitogo et al., 1998; Sung et al., 2001;
Zhang et al., 2016a, 2016b). So research of the structures of bacterial
communities had aroused interest in aquaculture. In Fig. 4, as the dom-
inant potential pathogens, the total relative abundance of
Piscirickettsiaceae and Flavobacteriaceae were 26.12%, 3.53%, 4.21% and
1.35% respectively in the control, BFT1, BFT2 and BFT3. Heterotrophic
bacteria played an indispensable role in controlling the water quality
in biofloc aquaculture systems, which inhibited the outbreak of the
pathogenic bacteria. The relative abundance of heterotrophic bacteria
in the top 30 dominant microbial communities at the family level in
BFT3 was higher than that of the control, BFT1 and BFT2 by 24.54%,
1.19% and 20.70%, respectively. Especially, the Rhodobacteraceae could
establish a beneficial bacterial community in the culture environment
of shrimp and thereby limit the survival of pathogenic bacteria, and
the relative abundance of Rhodobacteraceae in BFT3 was the highest
among the biofloc samples. So BFT3 was particularly favorable to the
cultivation of heterotrophic bacteria. Compared with sucrose, cassava
dregs can provide more attachment sites for heterotrophic bacteria,
which consumed hydrocarbon and nitrogen compounds. Furthermore,
enzyme-hydrolyzed cassava dregs have more attachment sites than su-
crose and higher simple sugar concentrations than original cassava
dregs.
As shown in Fig. 5, the ordination results showed that COD had the
largest impact on bacterial populations in L. vannamei aquaculture
water, followed by TSS and NO− 3 –N, further indicating that different
water quality factors had different influences on bacterial populations.
Compared with diversity indices like Shannon index and Simpson
index, biofloc had a greater bacterial diversity than shrimp intestines.
Johnson et al. reported that aquaculture water had substantially more
diverse microflora than the intestine of shrimp, because the environ-
ment of the shrimp intestines were different from the water in aquacul-
ture tanks, and not all microbes in the biofloc were ingested and survive
in the intestines(Moreira et al., 2006). Moreover, shrimp intestines
were less aerobic than tank water, and strictly anaerobic bacteria, such
as Propionigenium and Fusibacter, might deferentially select against cer-
tain bacteria were observed with immunological factors (Cardona et al.,
 Q. Shang et al. / Science of the Total Environment 615 (2018) 681–690
2016). However, shrimp consumed the biofloc, and thus, the microor-
ganisms in these bioflocs still affected the indigenous microflora in the
shrimp intestines (Cardona et al., 2016; Gatesoupe, 1999). As shown
in Fig. 7, the profiles of intestinal bacterial communities from each
group were distinct, and the correlation coefficients of the relative
abundance and variety of all microbial communities at the genus level
in the intestine samples (S0, S1, S2 and S3) and their corresponding
biofloc samples (the control, BFT1, BFT2 and BFT3) were 0.171, 0.305,
0.076 and 0.317, respectively. The distribution of microbes in the aquat-
ic environment influenced colony formation in the intestines of the
shrimp, especially in the BFT3 and S3 groups.
As the marine shrimp pathogen can be transmitted to shrimp
through water, as shown in Fig.7, Photobacterium damselae had a con-
siderable percentage in S0 and S2, while S3 had the lowest percentage
among the four samples, herein, it was an impositive correlation be-
tween the percentage of Photobacterium damselae in intestines with
the relative abundance of heterotrophic bacteria in correspondance
biofloc, implying that enzyme-hydrolyzed cassava dregs had greater ad-
vantages in inhibiting the survival of pathogens in shrimp intestines
than other carbon sources, especially than sucrose.
In summary, enzyme-hydrolyzed cassava dregs, the carbon source in
biofloc systems, had greatly promoted the formation of biofloc, which
were found to favor water quality control, microorganism reproduction,
and shrimp growth. Compared to cassava dregs, enzyme-hydrolyzed
cassava dregs had a higher simple sugar content, allowing for easy and
efficient absorption by heterotrophic bacteria. Besides, enzyme-
hydrolyzed cassava dregs providing more attachment sites for hetero-
trophic bacteria and other microorganism compared to sucrose and
resulting in faster ammonia removal, which maintained safe levels of ni-
trogen compounds for shrimp growth. Dense, active heterotrophic bac-
teria that dominate the population of aquatic microorganisms were
cultivated to inhibit the growth of pathogenic bacteria, thereby decreas-
ing the risk of shrimp diseases and improving shrimp survival rate.
More importantly, nutritious bioflocs were available in situ as a supple-
mental food source for the cultured shrimp, on one hand. Biofloc, as a
supplemental shrimp food source, has two potential advantages: It
could reduce the cost of shrimp feed, and its consumption and regener-
ation could recycle a fraction of the nutrients excreted by shrimp
(Hargreaves, 2006).
5. Conclusions
The study confirmed the feasibility of using cassava dregs as a substi-
tute for sucrose applied in biofloc aquaculture systems. The biofloc cul-
tivated by enzyme-hydrolyzed cassava dregs had obvious effect on the
controlling of water quality, better growth performance and higher
survival rate than sucrose and cassava dregs. Moreover, enzyme-
hydrolyzed cassava dregs were more easily assimilated by
heterotrophic bacteria and had more attachment sites for microorgan-
ism propagation. More importantly, applying enzyme-hydrolyzed cas-
sava dregs as the carbon source to the biofloc aquaculture system was
useful for the nonpoint source pollution reduction from random dis-
charge of cassava dregs. However, further studies still need to optimize
the method of enzymatic hydrolysis of cassava dregs to use this material
in biofloc aquaculture systems more effectively and economically.
Acknowledgements
This work was supported from the National Natural Science Founda-
tion of China (Nos. 41673105 and 41473118), and the Plan of Guangxi Sci-
ence Studies and Technology Development (Guikehe 14125007-2-13).
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.scitotenv.2017.08.256.
689
References
Agj, T., Cody, J.J., Conquest, L.D., et al., 2015. Effect of culture system on the nutrition and
growth performance of Pacific white shrimp Litopenaeus vannamei (Boone) fed dif-
ferent diets. Aquac. Nutr. 8 (2), 121–137.
Avnimelech, Y., 1999. Carbon/nitrogen ratio as a control element in aquaculture systems.
Aquaculture 176, 227–235.
Avnimelech, 2006. Bio-filters: the need for a new comprehensive approach. Aquac. Eng.
34, 172–178.
Avnimelech, Y., Kochba, M., 2009. Evaluation of nitrogen uptake and excretion by tilapia
in bio floc tanks, using 15N tracing. Aquaculture 287, 163–168.
Avnimelech, Y., Ritvo, G., 2003. Shrimp and fish pond soils: processes and management.
Aquaculture 220, 549–567.
Azim, M.E., Little, D.C., 2008. The biofloc technology (BFT) in indoor tanks: water quality,
biofloc composition, and growth and welfare of Nile tilapia (Oreochromis niloticus).
Aquaculture 283, 29–35.
Azim, M.E., Little, D.C., Bron, J.E., 2008. Microbial protein production in activated suspen-
sion tanks manipulating. Bioresour. Technol. 99, 3590.
Bowman, J.S., Rasmussen, S., Blom, N., Deming, J.W., Rysgaard, S., Sicheritz-Ponten, T.,
2012. Microbial community structure of Arctic multiyear sea ice and surface seawater
by 454 sequencing of the 16S RNA gene. ISME J. 6, 11.
Braak, T., Smilauer, P., 2002. Canoco reference manual and CanoDraw for Windows user's
guide: software for canonical community ord. Ithaca NY USA Www.
Burford, M.A., Thompson, P.J., Mcintosh, R.P., et al., 2004. The contribution of flocculated
material to shrimp (Litopenaeus vannamei) nutrition in a high-intensity, zero-
exchange system. Aquaculture 232 (1–4), 525–537.
Cardona, E., Gueguen, Y., Magré, K., Lorgeoux, B., Piquemal, D., Pierrat, F., et al., 2016. Bac-
terial community characterization of water and intestine of the shrimp Litopenaeus
stylirostris in a biofloc system. BMC Microbiol. 16, 157.
Chaiyapechara, S., Rungrassamee, W., Suriyachay, I., Kuncharin, Y., Klanchui, A.,
Karoonuthaisiri, N., et al., 2012. Bacterial community associated with the intestinal
tract of P. monodon in commercial farms. Microb. Ecol. 63, 938–953.
Chao, A., 1984. Nonparametric estimation of the number of classes in a population. Scand.
J. Stat. 11 (4), 265–270.
Clescerl, L.S., 1998. Standard methods for the examination of water and wastewater. 20th
Ed. American Public Health Association.
Crystaln, J., Sue, B., John, O., et al., 2010. Microbial community analysis of water, foregut,
and hindgut during growth of pacific white shrimp, Litopenaeus vannamei, in
closed-system aquaculture. J. World Aquacult. Soc. 39 (2), 251–258.
Daims, H., Brühl, A., Amann, R., et al., 1999. The domain-specific probe EUB338 is insuffi-
cient for the detection of all Bacteria: development and evaluation of a more compre-
hensive probe set. Syst. Appl. Microbiol. 22 (3), 434.
Das, S., Ward, L.R., Burke, C., 2008. Prospects of using marine actinobacteria as probiotics
in aquaculture. Appl. Microbiol. Biotechnol. 81, 419–429.
Ebeling, J.M., Timmons, M.B., Bisogni, J.J., 2006. Engineering analysis of the stoichiometry
of photoautotrophic, autotrophic, and heterotrophic removal of ammonia-nitrogen in
aquaculture systems. Aquaculture 257, 346–358.
Emerenciano, M., Elc, B., Cavalli, R.O., Wasielesky, W., 2012. Biofloc technology application
as a food source in a limited water exchange nursery system for pink shrimp
Farfantepenaeus brasiliensis (Latreille, 1817). Aquac. Res. 43, 447–457.
Fiorda, F.A., Soares Jr., M.S., Silva, F.A.D., et al., 2015. Physical quality of snacks and techno-
logical properties of pre-gelatinized flours formulated with cassava starch and
dehydrated cassava bagasse as a function of extrusion variables. LWT Food Sci.
Technol. 62 (2), 1112–1119.
Gaona, C.A.P., Poersch, L.H., Krummenauer, D., Fóes, G.K., Junior, W.W., 2011. The Effect of
Solids Removal on Water Quality, Growth and Survival of Litopenaeus vannamei in a
Biofloc Technology Culture System. 12 pp. 54–73.
Gatesoupe, F.J., 1999. The use of probiotics in aquaculture. Aquaculture 180,
147–165.
Hargreaves, J.A., 2006. Photosynthetic suspended-growth systems in aquaculture. Aquac.
Eng. 34, 344–363.
Hari, B., Kurup, B.M., Varghese, J.T., Schrama, J.W., Verdegem, M.C.J., 2006. The effect of
carbohydrate addition on water quality and the nitrogen budget in extensive shrimp
culture systems. Aquaculture 252, 248–263.
Haslun, J., Correia, E., Strychar, K., Morris, T., Samocha, T., 2012. Characterization of
bioflocs in a no water exchange super-intensive system for the production of food
size pacific white shrimp Litopenaeus vannamei. International J. Aquac.
Hopkins, J.S., Sandifer, P.A., Browdy, C.L., 1994. Sludge management in intensive pond cul-
ture of shrimp: effect of management regime on water quality, sludge characteristics,
nitrogen extinction, and shrimp production. Aquac. Eng. 13, 11–30.
Izquierdo, M., Forster, I., Divakaran, S., et al., 2006. Effect of green and clear water and lipid
source on survival, growth and biochemical composition of Pacific white shrimp
Litopenaeus vannamei. Aquac. Nutr. 12 (3), 192–202.
Ju, Z.Y., Ian, F., Lytha, C., et al., 2008. Determination of microbial community structures of
shrimp floc cultures by biomarkers and analysis of floc amino acid profiles. Aquac.
Res. 39 (2), 118–133.
Kirchman, D.L., 2002. The ecology of Cytophaga-Flavobacteria in aquatic environments.
FEMS Microbiol. Ecol. 39, 91–100.
Krummenauer, D., Poersch, L.H., Fóes, G., Lara, G., Wasielesky, W., 2016. Survival and
growth of Litopenaeus vannamei reared in Bft System under different water depths.
Aquaculture 465, 94–99.
Lavilla-Pitogo, C.R., Leaño, E.M., Paner, M.G., 1998. Mortalities of pond-cultured juvenile
shrimp, Penaeus monodon, associated with dominance of luminescent vibrios, in the
rearing environment. Aquaculture 164 (1–4), 337–349.
Martins, P., Cleary, D.F., Pires, A.C., Rodrigues, A.M., Quintino, V., Calado, R., et al., 2013.
Molecular analysis of bacterial communities and detection of potential pathogens
690 Q. Shang et al. / Science of the Total Environment 615 (2018) 681–690
in a recirculating aquaculture system for Scophthalmus maximus and Solea
senegalensis. PLoS One 8, 995–998.
Miura, Y., Hiraiwa, M.N., Ito, T., et al., 2007. Bacterial community structures in MBRs
treating municipal wastewater: relationship between community stability and reac-
tor performance. Water Res. 41 (3), 627–637.
Mooyoung, M., Anderson, W.A., Chakrabarty, A.M., 2001. Environmental biotechnology:
principles and applications. Environmental Biotechnology Principles & Applications.
Moreira, D., Rodríguez-Valera, F., López-García, P., 2006. Metagenomic analysis of meso-
pelagic Antarctic plankton reveals a novel delta proteobacterial group. Microbiology
152, 505–517.
Naylor, R.L., Goldburg, R.J., Mooney, H., Beveridge, M., Clay, J., Folke, C., et al., 1998.
Nature's subsidies to shrimp and salmon farming. Science 282, 883–884.
Payne, M.S., Hall, M.R., Sly, L., et al., 2007. Microbial diversity within early-stage cultured
Panulirus ornatus phyllosomas. Appl. Environ. Microbiol. 73 (6), 1940–1951.
Pitta, D.W., Pinchak, W.E., Dowd, S.E., Osterstock, J., Gontcharova, V., Youn, E., et al., 2010.
Rumen bacterial diversity dynamics associated with changing from bermudagrass
hay to grazed winter wheat diets. Microb. Ecol. 59, 511–522.
Ramette, A., 2007. Multivariate analyses in microbial ecology. FEMS Microbiol. Ecol. 62,
142–160.
Rungrassamee, W., Klanchui, A., Chaiyapechara, S., Maibunkaew, S., Tangphatsornruang,
S., Jiravanichpaisal, P., et al., 2013. Bacterial population in intestines of the black
tiger shrimp (Penaeus monodon) under different growth stages. PLoS One 8, e60802.
Sakami, T., Fujioka, Y., Shimoda, T., 2008. Comparison of microbial community structures
in intensive and extensive shrimp culture ponds and a mangrove area in Thailand.
Fish. Sci. 74 (4), 889–898.
Schryver, P.D., Crab, R., Defoirdt, T., Boon, N., Verstraete, W., 2008. The basics of bio-flocs
technology: the added value for aquaculture. Aquaculture 277, 125–137.
Schveitzer, R., Arantes, R., Pfs, C., Cm, D.E.S., Arana, L.V., Seiffert, W.Q., et al., 2013. Effect of
different biofloc levels on microbial activity, water quality and performance of
Litopenaeus vannamei in a tank system operated with no water exchange. Aquac.
Eng. 56, 59–70.
Shang, Y.C., Leung, P., Ling, B.H., 1998. Comparative economics of shrimp farming in Asia.
Aquaculture 164, 183–200.
Shannon, C.E.A., 2001. A mathematical theory of communication. AT&T Tech J. Acm
Sigmobile Mobile Computing & Communications Review 5 (1), 3–55.
Singh, K.M., Tripathi, A.K., Pandya, P.R., et al., 2010. Methanogen diversity in the rumen of
Indian Surti buffalo (Bubalus bubalis), assessed by 16S rDNA analysis. Res. Vet. Sci. 51
(3), 451–455.
Steicke, C., Jegatheesan, V., Zeng, C., 2007. Mechanical mode floating medium filters for
recirculating systems in aquaculture for higher solids retention and lower freshwater
usage. Bioresour. Technol. 98, 3375–3383.
Sun, L., Wan, S., Yu, Z., Wang, Y., Wang, S., 2012. Anaerobic biological treatment of high
strength cassava starch wastewater in a new type up-flow multistage anaerobic reac-
tor. Bioresour. Technol. 104, 280.
Sung, H.H., Hsu, S.F., Chen, C.K., et al., 2001. Relationships between disease outbreak in
cultured tiger shrimp (Penaeus monodon) and the composition of Vibrio communities
in pond water and shrimp hepatopancreas during cultivation. Aquaculture 192 (2),
101–110.
Tomoko, S., Yoshimi, F., Toru, S., 2010. Comparison of microbial community structures in
intensive and extensive shrimp culture ponds and a mangrove area in Thailand. Fish.
Sci. 74 (4), 889–898.
Van Wyk, P., Davishodgkins, M., Laramore, R., Main, K.L., Mountain, J., Scarpa, J., 1999.
Farming Marine Shrimp in Recirculating Freshwater Systems.
Ventura, M., Canchaya, C., Tauch, A., Chandra, G., Fitzgerald, G.F., Chater, K.F., et al., 2007.
Genomics of Actinobacteria: tracing the evolutionary history of an ancient phylum.
Microbiology & Molecular Biology Reviews Mmbr 71, 495–548.
Wagner, M., Amann, R., Lemmer, H., et al., 1993. Probing activated sludge with oligonucle-
otides specific for proteobacteria: inadequacy of culture-dependent methods for de-
scribing microbial community structure. Appl. Environ. Microbiol. 59 (5), 1520–1525.
Wagner, M., Rath, G., Amann, R., et al., 1995. In situ, identification of ammonia-oxidizing
bacteria. Syst. Appl. Microbiol. 18 (2), 251–264.
Wang, G., Yu, E., Xie, J., Yu, D., Li, Z., Luo, W., et al., 2015. Effect of C/N ratio on water quality
in zero-water exchange tanks and the biofloc supplementation in feed on the growth
performance of crucian carp, Carassius auratus. Aquaculture 443, 98–104.
Wasielesky Jr., Wilson, Atwood, H., Stokes, A., et al., 2006. Effect of natural production in a
zero exchange suspended microbial floc based super-intensive culture system for
white shrimp Litopenaeus vannamei. Aquaculture 258 (1–4), 396–403.
Wei, Y.F., Liao, S.A., Wang, A.L., 2016. The effect of different carbon sources on the nutri-
tional composition, microbial community and structure of bioflocs. Aquaculture
465, 88–93.
Wickins, J.F., 1976. The tolerance of warm-water prawns to recirculated water. Aquacul-
ture 9, 19–37.
Woebken, D., Fuchs, B.M., Kuypers, M.M., Amann, R., 2007. Potential interactions of
particle-associated anammox bacteria with bacterial and archaeal partners in the
Namibian upwelling system. Appl. Environ. Microbiol. 73, 4648.
Xu, W.J., Pan, L.Q., 2012. Effects of bioflocs on growth performance, digestive enzyme ac-
tivity and body composition of juvenile Litopenaeus vannamei in zero-water exchange
tanks manipulating C/N ratio in feed. Aquaculture s356–357 (4), 147–152.
Yoram, A., Malka, K., 2009. Evaluation of nitrogen uptake and excretion by tilapia in bio
floc tanks, using 15N tracing. Aquaculture 287, 163–168.
Yu, Y., Wei, J., Zhang, X., Liu, J., Liu, C., Li, F., et al., 2014. SNP discovery in the transcriptome
of white Pacific shrimp Litopenaeus vannamei by next generation sequencing. PLoS
One 9, e87218.
Zeng, Y.H., Ma, Y., Wei, C.L., et al., 2010. Bacterial diversity in various coastal mariculture
ponds in Southeast China and in diseased eels as revealed by culture and culture-
independent molecular techniques. Aquac. Res. 41 (9), e172–e186.
Zhang, H., Sun, Z., Liu, B., Xuan, Y., Jiang, M., Pan, Y., et al., 2016a. Dynamic changes of mi-
crobial communities in Litopenaeus vannamei cultures and the effects of environmen-
tal factors. Aquaculture 455, 97–108.
Zhang, M., Xie, L., Yin, Z., Khanal, S.K., Zhou, Q., 2016b. Biorefinery approach for cassava-
based industrial wastes: current status and opportunities. Bioresour. Technol. 215,
50–62.
Zhou, A.J., CY, Fu, SY, Fu, 2014. The pattern of enzymatic hydrolysis of starch and cellulose
in cassava dregs. Paper Science & Technology.