1. Introduction
Monitoring the health and migration of marine organisms holds immense importance to humanity. It plays a key role in maintaining the equilibrium of marine ecosystems, advancing climate change studies, and safeguarding human health and safety, as well as preserving biodiversity
[1],
[2],
[3],
[4]. Therefore, this has driven the rapid development of wearable sensors for aquatic animals
[5], such as utilizing wearable resistive bending sensors to monitor underwater animal speeds
[6]. Sensors designed for underwater organisms must exhibit strong adhesion and biocompatibility. Unlike humans, marine organisms have smoother, more hydrated skin and often undergo rapid and significant deformations to dislodge attachments
[7],
[8]. Furthermore, the marine environment impacts adhesive bonding and accelerates substance diffusion. Thus, developing flexible devices with strong adhesion and excellent biocompatibility is the key parameter in marine organism sensing technology, ensuring that sensors can adhere firmly to skin or other surfaces to obtain reliable and stable signals. Additionally, high sensitivity, evaluated by the gauge factor (GF), which is defined as the ratio of the relative change in the output signal to the applied strain, is essential
[9]. It is crucial for monitoring biological signals such as heartbeat, respiration, and movement, as it determines the device’s ability to detect subtle physiological changes
[10].
The marine ecological environment is complex and directly affects the activity levels and metabolic rates of marine organisms. For instance, organisms like scallops, which have low metabolic rates and less distinct cardiac structures, display slow and minimal heart rate variations, requiring smaller strain responses in corresponding sensors
[11],
[12],
[13]. Thus, developing sensors for marine organisms with broader applications requires higher GF values. A range of materials has been explored for detecting marine organisms
[14],
[15],
[16],
[17],
[18],
[19],
[20]. Among these materials, hydrogels demonstrate good stretchability and possess soft, moist properties, making them highly promising for applications in flexible electronics
[21],
[22],
[23]. To enhance biocompatibility, improve adhesion, and reduce cytotoxicity of hydrogels for sensor applications, natural proteins such as silk fibroin
[24],
[25],
[26], collagen
[27], bovine serum albumin
[28], and gluten
[29] are commonly incorporated as fundamental components in constructing the hydrogel network. However, natural protein hydrogel sensors often exhibit low conductivity, poor mechanical strength, and insufficient self-recovery capabilities
[30]. Some efforts have been made to address these challenges; for example, Roshanbinfar et al.
[31] incorporated gold nanoparticles (AuNPs) to enhance the conductivity of collagen hydrogels. Xu et al.
[32] aimed to enhance mechanical strength by utilizing interactions between acrylic acid (AA) and acrylamide (AM) to form cross-linked networks. Enhancing the structural stability and mechanical performance of protein hydrogels often requires increased cross-linking, which may reduce sensitivity
[33],
[34],
[35]. Additionally, conductive fillers (carbon nanotubes, graphene, and silver nanowires) are prone to aggregation in the protein hydrogels
[36],
[37].
To overcome these challenges, we designed a novel sensing material that combines a natural protein (keratin, KE) with fluid liquid metal (LM). KE is a protein widely distributed in tissues such as feathers and wool (
Fig. 1(a)). It primarily consists of sulfur-containing amino acids, such as cysteine and methionine, which enable cross-linking via hydrogen and disulfide bonds. It also has abundant free thiol groups
[38],
[39],
[40] that can form metal-thiol coordination interactions
[41]. Leveraging these interactions, we introduced LM as a conductive filler to enhance the conductivity of protein hydrogels. Eutectic gallium indium (EGaIn), known for its deformability and conductivity, was used to improve the mechanical properties of soft materials
[42]. Its interaction with thiol ligands allows EGaIn to be incorporated into the protein network, further enhancing both mechanical and electrical properties
[43],
[44],
[45].
Therefore, we selected KE and LM for their flexibility and deformability, achieving effective binding of LM with KE through thiol groups. This LM/KE composite solution was used to fabricate the keratin liquid metal (KELM) hydrogel, addressing the issues of low mechanical strength and poor conductivity in protein hydrogels. Additionally, the metastability of protein secondary structures allows the α-helix and β-sheet configurations of KE to undergo reversible or irreversible transitions, imparting shape memory characteristics
[46]. Our study revealed that the KELM hydrogel exhibits high biocompatibility and minimal cytotoxicity, as evidenced by the proliferation assay of human epidermal keratinocyte (HACAT) cell lines. This approach not only effectively overcomes the common issue of conductive filler aggregation but also integrates shape memory functionality. Consequently, the KELM hydrogel can monitor subtle heartbeat movements in invertebrate aquatic animals and detect heartbeat and continuous tail movements in aquatic vertebrates. Additionally, it can differentiate between the heartbeat patterns of various aquatic organisms, such as giant salamanders, sturgeon, and scallops, providing a robust method for the timely and accurate assessment of aquatic health.
2. Materials and methods
2.1. Materials
KE was purchased from Xi’an Muguo Biotechnology (China). LM (EGaIn with a content of 75.5 wt% gallium (Ga)/24.5 wt% indium (In), ≥ 99.99%), AM (≥ 99%), N,N′-methylenebis acrylamide (BIS, 99%), ammonium persulfate (APS, ≥ 98%), and N,N,N′,N′-tetramethyl ethylenediamine (TEMED, 99%) were purchased from Sigma Aldrich (USA).
2.2. Preparation of the KELM hydrogel
The KELM hydrogel was prepared by dispersing LM in a 5% KE solution, followed by 30 min of ultrasonication in an ice bath to obtain a homogeneous LM/KE suspension. A mixture of 2 mL of 50 wt% AM, 500 µL of 0.2 mol∙L−1 APS, 350 µL of 0.2 wt% BIS, and 10 µL of TEMED was quickly mixed with 1.2 mL of the LM/KE solution, and then injected into the mold. The resulting KELM hydrogel was stored in a refrigerator at 4 °C. The KELM hydrogels were prepared with different mass ratios of LM to KE, represented as 0.10, 0.15, 0.20, 0.25, and 0.30. The detailed formulations are provided in Table S1 in Appendix A. For comparison, the polyacrylamide hydrogel (PAM) was prepared as a blank control without the addition of the LM/KE suspension, and the KE hydrogel was prepared without LM using the same method.
2.3. Characterization
Scanning electron microscope (SEM; ZEISS sigma500; Zeiss, Germany), transmission electron microscopy (TEM; JEM-2100F; JEOL, Japan), tensile testing machine (ZQ-990L; Zhiqu Instrument, China), inductance, capacitance, and resistance (LCR) meter (E4980AL; Keysight, USA), X-ray photoelectron spectroscopy (XPS; Thermo Scientific, USA), fourier transform infrared (FTIR) spectrophotometer (Shimadzu, Japan), laser scanning confocal microscope (STELLARIS STED; Leica, Germany), ultrasonic cleaner (PS-20A; Xinzhi, China), dynamic light scattering (DLS; Zetasizer Nano ZS90; Malvern Instruments, UK), vacuum freeze dryer (X0-18S; Xianou, China), flow cytometry (F1325; Beckman, USA), microplate reader (SuPerMax 3100; Flash, China), and smartphone (P60; Huawei, China) were used in this work.
2.4. Mechanical testing
The tensile properties of the hydrogels were determined using a mechanical testing machine at room temperature. The hydrogels were 30 mm long, 10 mm wide, and 2 mm thick, with a stretching speed of 100 mm∙min−1. The stability of the hydrogel was tested through 100 repeated cycles.
2.5. Electrical testing
Changes in resistance signals under different strains, compressions, and fatigue resistance were evaluated using a mechanical testing machine and an LCR instrument. The hydrogel’s response time and electrical performance were assessed, and conductivity (S∙m−1) was calculated using the following formula:
where σ (S∙m−1) represents conductivity, L (m) represents the distance between the test electrodes, R (Ω) is the resistance of the hydrogel, and S (m2) is the cross-sectional area of the hydrogel. The sensitivity of hydrogels is typically expressed by the GF, which is calculated using the following equation:
where R0 is the initial resistance, ΔR is the change in resistance (ΔR = R−R0), and ε is the applied strain.
2.6. Adhesion testing
Tensile bonding tests were conducted on hydrogels of uniform size (2 mm thick, 15 mm × 15 mm in area) to measure the adhesion strength on various substrates. The tests were performed in tension mode at a speed of 300 mm∙min−1, with polyethylene terephthalate (PET), polyimide (PI), rubber, and pigskin cut into rectangular sections measuring 40 mm × 20 mm. The δ was calculated as:
where δ (kPa) is shear strength, and F (N) is the maximum shear force.
2.7. Biocompatibility testing
To evaluate the cytotoxicity of the hydrogels toward HACAT cells, the cell viability and biocompatibility were assessed by the Cell counting kit-8 (CCK-8) method. A hydrogel with a diameter of 14 mm was placed at the bottom of a 24-well plate, and cells were seeded at a density of 1 × 105 cells∙well−1. Cells cultured on a slide served as the control group. After 24 h of incubation, the culture medium was replaced with fresh medium containing 10% CCK-8 reagent. The cells were then incubated for an additional 3 h at 37 °C. The medium from each well (100 μL per well) was transferred to a 96-well plate, and the absorbance at 450 nm was recorded using a microplate reader. Cell apoptosis was evaluated by flow cytometry. Collected cells were stained with fluorescein isothiocyanate (FITC)-annexin V and PI using an apoptosis detection kit (40302ES60, Yeasen, China). Early apoptosis (FITC-annexin V+/PI− cells) and late apoptosis were measured using flow cytometry.
Cell morphology was observed to further assess biocompatibility. After a culture for 5 days, cells were fixed with 4% paraformaldehyde for 30 min and permeabilized with 0.1% Triton X-100 for 15 min. And then, cells were blocked with 3% bovine serum albumin (BSA) for 30 min and stained with FITC-conjugated wheat germ agglutinin (WGA) to visualize the cell membrane. Cell nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI). The stained cells were observed under a laser-scanning confocal microscope.
3. Results and discussion
3.1. Preparation of KE dispersed LM suspension and the KELM hydrogel
Fig. 1(a) illustrates the process of mixing LM with a 5% KE solution, followed by ultrasonication. This procedure resulted in a uniformly dispersed LM/KE solution, as shown in Fig. S1 in Appendix A. Several dispersants, including the commercial surfactant sodium dodecyl sulfate (SDS)
[47] and hyaluronic acid (HA)
[48], were evaluated for their effectiveness in dispersing LM (Fig. S2 in Appendix A). Following 24 h of ultrasonication, LM nanoparticles aggregated and precipitated at the bottom of the SDS and HA dispersions. In contrast, the LM/KE solution remained relatively uniform, demonstrating superior dispersibility of KE-dispersed LM compared to SDS and HA. LM/KE nanoparticles were used as a conductive filler in the high-performance KELM hydrogel, incorporating thiol-metal coordination and hydrogen bonds into the network (
Fig. 1(b)). Thiolated ligands bound effectively to the EGaIn interface, and TEM images revealed KE attached to LM particles (
Figs. 1(c-i) and (c-ii)). Elemental mapping confirmed the presence of Ga from LM and sulfur (S) from KE (Fig. S3 in Appendix A). Particle size analysis showed LM/KE had a larger size than LM (Fig. S4(a) in Appendix A). The zeta potential of LM was −26.4 mV, whereas that of LM/KE was −2.16 mV, indicating positive KE binding to the LM surface (Fig. S4(b) in Appendix A). These results confirm the successful coordination of thiolated KE with LM.
XPS of the S 2
p region confirmed the coordination of –SH groups with EGaIn, showing S–Ga bonds at 160.8 and 161.18 eV
[49] in the KELM hydrogel, compared to 162.84 and 164.14 eV in the KE hydrogel, which indicated successful EGaIn–SH bonding (
Fig. 1(d) and Fig. S5 in Appendix A). As shown in Fig. S6 in Appendix A, the elemental mapping of the KELM hydrogel indicated KE binding to the LM surface by –SH groups. Additionally, FTIR showed that stretching vibration peaked at 3346 and 3181 cm
−1, indicating hydrogen bond formation in the KELM hydrogel (Fig. S7 in Appendix A). The integration of EGaIn into the hydrogel network enhanced its mechanical and electrical properties, achieving a high GF of 7.03, surpassing that of previously reported hydrogels with silk fibroin, ferritin, gluten, and antifreeze protein (
Fig. 1(e), Table S2 in Appendix A).
3.2. Mechanical and adhesive properties of the KELM hydrogel
The prepared KELM hydrogel exhibited high plasticity, allowing for the molding of intricate patterns (
Fig. 2(a)). A high-water content is crucial for preserving the mechanical flexibility of hydrogels. Various KELM hydrogels maintained approximately 70% water content (
Fig. 2(b)). Due to its high-water content and the inclusion of LM/KE, the hydrogel exhibited ultra-soft properties and could be easily stretched by hand (
Fig. 2(c)). The enhancement of the mechanical properties of the KELM hydrogel can be explained by the following mechanisms (
Fig. 2(d)): stretching caused the KE chains to extend and elongate
[46], increasing the hydrogel’s ability to deform under external forces. Simultaneously, the attached LM nanoparticles, known for their high flexibility and stretchability, contributed to the hydrogel’s softness and elongation.
By adjusting the mass ratio of LM to KE, the KELM hydrogels were designed to achieve significant improvements in toughness, which were further evaluated through tensile tests. As shown in
Fig. 2(e), the KE hydrogel exhibited superior tensile performance compared to the PAM hydrogel, owing to the abundance of hydrogen bonds in KE that enhanced its mechanical properties. Increasing the EGaIn content further improved the mechanical performance, with the stretchability of the KELM hydrogel rising progressively with the LM nanoparticle-to-protein ratio (Fig. S8 in Appendix A). At an LM-to-KE mass ratio of 0.30, the maximum tensile stress and elongation at break increased from 92.29 kPa and 1764% to 165.89 kPa and 2567%, respectively. Additionally, ten consecutive loading–unloading cycles were conducted on the KELM hydrogel at 100% strain (
Fig. 2(f)). The KELM hydrogel demonstrated consistent hysteresis loops and stable mechanical properties throughout the stretching and releasing cycles.
Another notable advantage of the KELM hydrogel is its robust adhesion capability. Improving the adhesion of wearable electronic devices to human skin or prostheses is crucial for conformal attachment of strain sensors, avoiding interface delamination under large deformations, and ensuring accurate monitoring. Tensile-shear adhesive tests assessed the adhesion strength of the KELM hydrogel on PI, PET, rubber, and pigskin, with measured strengths of approximately 14.61, 17.23, 37.18, and 5.16 kPa, respectively (
Fig. 2(g)). This good adhesion performance was evident across diverse substrates, including organic materials (PVC and wood), inorganic materials (aluminum and glass), and biological surfaces (shells and snail shells) (Fig. S9 in Appendix A). The hydrogel exhibited excellent underwater adhesion, securely holding objects weighing approximately 50 g (
Fig. 2(h)) and maintaining stable attachment without slipping (Video S1 in Appendix A). This strong adhesive property arises from hydrolyzed KE’s abundance in amino acids, which provide numerous amino and carboxyl groups for bonding through physical interactions such as hydrogen bonding and metal chelation
[50]. Consequently, the KELM hydrogel is soft, stretchable, and mechanically stable, closely resembling natural skin, making it highly suitable for wearable sensor applications.
3.3. Biocompatibility of the KELM hydrogel
In addition to the essential properties of stretchability and adhesion, a high-performance wearable sensor must exhibit biocompatibility. The KELM hydrogel’s biocompatibility is attributed to its KE protein. As shown in
Fig. 3(a), the KELM hydrogel adhered well to wrist skin without causing adverse reactions such as skin allergies. Application of it to aquatic animals, like sturgeon (
Fig. 3(b)), also showed no residue, swelling, or allergic responses.
To further assess biocompatibility, cytotoxicity tests were conducted on HACAT cells, evaluating cell proliferation, viability, and apoptosis.
Fig. 3(c) demonstrated that cell proliferation in the experimental group was comparable to the control group, indicating the non-toxic nature of the KELM hydrogel. This biocompatibility is primarily due to the KE protein. Additionally, cell apoptosis analysis (
Figs. 3(d) and (e)) revealed that over 90% of cells were viable (lower left quadrant). Although early apoptotic (lower right quadrant), late apoptotic (upper left quadrant), and necrotic (upper right quadrant) cells were present, their proportions were minimal and similar to the control group, confirming the hydrogel’s excellent cell compatibility.
For a more intuitive evaluation of cell survival, immunofluorescence staining was performed on HACAT cells cultured on the hydrogel (
Fig. 3(f)). The cells on the KELM hydrogel displayed normal morphology (green) and intact nuclear structure (blue), comparable to the control group, indicating excellent biocompatibility. These results strongly suggest that the KELM hydrogel exhibits great cell compatibility, making it a promising candidate for electronic skin, implantable devices, and other biomedical applications.
3.4. Sensing performances of the KELM hydrogel sensor
The electrical properties of hydrogels are crucial for their performance as strain sensors. The incorporation of LM significantly enhances the conductivity of the KELM hydrogel. As shown in Fig. S10 in Appendix A , the KELM-hydrogel-powered light emitting diode (LED) bulbs exhibited greater brightness compared to those powered by the KE hydrogel, reflecting improved conductivity due to the incorporation of LM (
Fig. 4(a)). Electrical conductivity increased with the LM-to-KE ratio, reaching 6.84 S∙m
−1 at a 0.30 ratio (Fig. S11 in Appendix A). Additionally, the LED brightness decreased as the hydrogel was stretched (Fig. S12 in Appendix A). The high conductivity endows KELM hydrogel-based strain sensors with exceptional sensitivity and reliability.
The strain sensitivity of the KELM hydrogel was assessed by monitoring resistance changes during stretching and releasing cycles. As shown in
Fig. 4(b), resistance increased during stretching and decreased during release without significant delay, indicating that the hydrogel’s electrical stability is due to its elastic network properties. The hydrogel demonstrated rapid response times of 240 ms for stretching and 200 ms for recovery (
Fig. 4(c)). In continuous stretching-release cycles, resistance increased linearly with strain (
Fig. 4(d)). Fig. S13 in Appendix A displayed the relative resistance change (Δ
R/
R0) and sensitivity of the hydrogel sensor across various strains. For strains between 0–150%, Δ
R/
R0 increased linearly with a GF of 2.50. Beyond 150% strain, Δ
R/
R0 rose sharply, with the GF increasing from 2.50 to 7.03 (150%–600%), indicating high sensitivity over a broad range. Given that the human dermis deforms between 0–75%, this hydrogel is well-suited for skin-like sensors to monitor human movements. The hydrogel sensor also showed excellent anti-fatigue performance, maintaining stable responses after 100 cycles of 50% strain (
Fig. 4(e)), confirming its stability and recoverable sensing capability.
The stability of the KELM hydrogel is further demonstrated by its shape memory properties. This shape memory effect is attributed to the inherent characteristics of KE protein molecules and the hydrogel’s structure. KE, a structurally complex protein, contains numerous α-helical structures within its molecular framework. These α-helices confer high flexibility and extensibility to KE, enabling reversible deformation under applied stress
[46]. The hydrogen bond network acts as a locking mechanism to stabilize the deformed shape of the hydrogel. During external stretching, KE’s α-helical structures unfold and transition into β-sheet formations. Heat exposure then fixes the hydrogel in this new stable configuration. Upon rehydration, the intermolecular hydrogen bonds in the β-sheets are broken, allowing KE molecules to revert to their original α-helical structures and restore the hydrogel to its initial shape (
Fig. 4(f)).
Wrapping the KELM hydrogel around a centrifuge tube and heating it until dry locks the hydrogel into a fixed spiral shape, resulting in a loss of flexibility. Spraying deionized water onto the hydrogel then rapidly restores it to its original stretchable and soft state (
Fig. 4(g)). Additionally, as shown in Fig. S14 in Appendix A , after stretching and drying the strip-shaped KELM hydrogel, the stretched shape was fixed. Immersion in deionized water caused the hydrogel to contract and nearly return to its original length (Video S2 and Fig. S15 in Appendix A). This shape memory property allows the KELM hydrogel to recover its original shape rapidly and automatically after external deformation. Shape memory hydrogels can be utilized to develop flexible sensors that conform to various surface topographies (e.g., fish skin) during observation. The hydrogel sensor’s ability to revert to its original shape following deformation due to external forces or environmental changes ensures the its prolonged stability, reusability, and consistent sensitivity.
3.5. Heartbeat monitoring of the KELM hydrogel sensor
Monitoring the heartbeat of aquatic animals is vital for evaluating their physiological status and overall health. Accurate and timely heartbeat monitoring is essential for health assessment and disease tracking in aquatic species. Leveraging its biomimetic mechanical properties, strong adhesion, stable electrical conductivity, and excellent sensitivity, the KELM hydrogel functions as a biomimetic skin sensor with significant potential for aquatic heartbeat monitoring.
As demonstrated in
Fig. 5(a), KELM hydrogel sensors were applied to invertebrates like scallops, vertebrates like sturgeons, and amphibians like giant salamanders. These sensors tracked heartbeat status and transmitted resistance signals wirelessly via Bluetooth to mobile phones or computers. Despite minor deformations (5.0%, 2.5%, and 1.0%), the sensors reliably produced normal signals (
Fig. 5(b)) and detected subtle changes during movement. The hydrogel’s soft structure and strong adhesion allowed it to conform well to the heart’s surface, ensuring stable attachment.
In Fig. S16 in Appendix A, a surgical tool was used to expose the anterior heart surface of a giant salamander. Despite this, the heart continued to beat (Video 3 in Appendix A). Strain sensors were placed on the heart’s surface and connected to an LCR meter for real-time monitoring (
Fig. 5(c)). The KELM hydrogel sensor was also connected to a multifunctional wireless printed circuit board to enable bluetooth transmission to mobile phones, allowing clear observation of sturgeon heartbeats (Video 4 in Appendix A). This setup provides high portability and flexibility, facilitating easy data acquisition, processing, and storage.
The KELM hydrogel’s relative resistance changes due to heartbeat-induced stretching enable real-time monitoring of various aquatic animals, including scallops, sturgeons, and giant salamanders (
Figs. 5(d)–(f)). Additionally, the KELM hydrogel sensor monitors significant tail swings in sturgeons (Fig. S17 in Appendix A) and various human movements (Fig. S18 in Appendix A), providing electronic feedback on limb movements such as wrist, finger, and knee bending. Consequently, biomimetic skin hydrogel sensors are promising wearable devices for monitoring aquatic animal movements, supporting timely medical diagnostics and personal health monitoring.
In our study, the heartbeat monitoring of mollusks, such as scallops, was conducted using a non-invasive method. This approach not only effectively collects physiological data but also reduces interference with the animals. However, for vertebrate aquatic animals, heartbeat monitoring requires invasive methods. The surgical procedures involved in monitoring may disrupt the physiological status of these animals. To address this issue, future research should focus on developing minimally invasive installation techniques, such as utilizing nanoscale or microscale devices, to minimize incisions and reduce interference with the biological subjects. Additionally, improving the postoperative recovery environment can significantly alleviate physiological stress in the animals, helping them return to their normal state more quickly. Furthermore, to achieve more precise behavioral monitoring, we will increase the dimensions of the monitoring by employing multiple sensor arrays to gather more detailed information. This multidimensional data collection approach will facilitate a more comprehensive understanding of the physiological states and behavioral patterns of aquatic animals. Moreover, subsequent data processing and analysis also need to be refined to enhance the accuracy and reliability of the data, thereby supporting more in-depth biological research.
4. Conclusions
In summary, by integrating cysteine-rich KE with LM and constructing a biocompatible hydrogel, we have effectively addressed the limitations of traditional protein hydrogels in mechanical performance and conductivity. The KELM hydrogel demonstrates good mechanical properties, including ultra-high tensile strength and an elongation at break of 2600%, while maintaining stability through tensile-release cycles. Its strong adhesion enables broad application to various substrates, including human skin and biological tissues. Notably, the skin-like hydrogel exhibits high sensitivity (GF = 7.03), durability, rapid response, and excellent shape memory across a wide range of pressures and strains, making it suitable for monitoring aquatic animal heartbeats. In conclusion, the KELM hydrogel combines superior mechanical and conductive properties with diverse application potential, offering promising opportunities for flexible electronics, medical monitoring, and wearable devices.
CRediT authorship contribution statement
Lidong Wu: Writing – review & editing, Writing – original draft, Supervision, Project administration, Methodology, Funding acquisition, Conceptualization. Jinxue Zhao: Writing – original draft, Methodology, Investigation, Formal analysis. Yuanxin Li: Formal analysis, Data curation. Haiyang Qin: Formal analysis, Data curation. Xuejing Zhai: Data curation. Peiyi Li: Data curation. Yang Li: Resources. Yingnan Liu: Resources. Ningyue Chen: Writing – review & editing. Yuan Li: Writing – review & editing.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (22176221 and 22273045), the Central Public-interest Scientific Institution Basal Research Fund, Chinese Academy of Fishery Sciences (2024XT09), the Tsinghua University Independent Scientific Research Plan for Young Investigator, and the Tsinghua University Initiative Scientific Research Program.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.eng.2024.12.030.
PDF
(2606KB)
