Biological valorization of Pangas fish processing waste into nutrient-rich fish silage using molasses-assisted Lactobacillus plantarum: fermentation dynamics, nutritional quality, and economic feasibility

Biological valorization of fish processing waste offers a sustainable strategy to reduce environmental load while generating value-added feed resources. This study evaluated the production of fish silage from Pangas (Pangasianodon hypophthalmus) processing waste using molasses-assisted Lactobacillus plantarum fermentation. A two-factor factorial design was employed with molasses levels (0–25% w/w) and L. plantarum inoculum levels (0–5% v/w). Fermentation was carried out for 30 days, and parameters including fermentation dynamics, proximate composition, biochemical stability, microbiological quality, and economic feasibility were assessed. The optimized treatment (15% molasses + 5% L. plantarum) yielded 0.89 kg silage per kg viscera and 0.37 kg per kg whole fish. Rapid acidification (pH < 4.5), decreasing temperature trends, and reduced water activity confirmed effective fermentation. Proximate composition showed improved crude protein, stabilized fat content, and progressive carbohydrate utilization. Free fatty acid and peroxide values remained within acceptable limits, indicating lipid stability. Total plate count ranged from 6.20 to 7.67 log CFU g -1, reflecting active fermentation, while Escherichia coli and Salmonella were not detected. Colour parameters remained stable across treatments. Economic analysis estimated a production cost of ₹93.70/kg at laboratory scale. Molasses supplementation enhanced fermentable substrate availability, promoting efficient lactic acid fermentation and stabilization of fish silage. The optimized combination of 15% molasses and 5% L. plantarum provided the best balance between fermentation efficiency, nutritional quality, and economic feasibility. The findings support the potential of this approach as a sustainable and cost-effective method for converting fish processing waste into aquafeed ingredients.

The rapid expansion of aquaculture has positioned it as the fastest-growing food production sector globally, contributing significantly to food security and livelihoods. In India, total fish production reached 17.54 million tons (MT) in 2022–23, with aquaculture accounting for nearly 75% of total output (Handbook of Fisheries Statistics, GOI, 2023). This growth has intensified the demand for formulated aquafeeds, which are projected to reach approximately 20 MT by 2025. Feed costs represent nearly 60% of total aquaculture production expenses, largely driven by the increasing prices and limited availability of conventional protein sources such as fishmeal and soybean meal (Sahu and Mohanta, 2023; Kantha and Prachoom 2024). Consequently, identifying sustainable, cost-effective, and locally available alternative protein ingredients has become a critical priority for the aquaculture sector.

Parallel to the expansion of aquaculture, the fisheries processing industry generates substantial quantities of organic waste. Postharvest losses in fisheries have remained persistently high, estimated at 20–35% of total production, while global fish waste generation exceeds 100 million tons annually, with India contributing more than 4 million tons (Binsi, 2018). Fish processing operations such as filleting produce 30–50% by-products in the form of heads, frames, viscera, skin, and fins (Ghaly et al., 2013). Although often discarded or underutilized, these by-products are rich in high-quality proteins, lipids, and minerals, making them suitable substrates for conversion into value-added feed ingredients.

Fish silage is a semi-liquid product obtained through the autolytic hydrolysis of fish tissues under acidic or microbial fermentation conditions and represents an effective approach for recycling fish processing waste into animal feed. Traditional acid silage methods rely on inorganic acids to reduce pH and stabilize the material; however, such approaches may result in protein denaturation, handling hazards, and residual acidity that limit feed quality and acceptance (Libonatti et al., 2023; Yilmaz et al., 2022). In contrast, biological silage produced through lactic acid bacteria (LAB) fermentation has gained increasing attention as a safer and more sustainable alternative, offering improved nutritional quality and microbiological stability.

Among LAB, Lactobacillus plantarum is widely employed in silage fermentation due to its homofermentative metabolism, adaptability to variable substrates, and ability to rapidly produce lactic acid, thereby lowering pH and suppressing spoilage and pathogenic microorganisms (Yilmaz et al., 2022). LAB fermentation has been shown to enhance protein hydrolysis, increase digestibility, and improve functional properties of silage, making it particularly suitable for aquafeed applications (Vidotti et al., 2002). However, successful fermentation depends on several factors, including substrate freshness, inoculum density, anaerobic conditions, and the availability of fermentable carbohydrates.

Fish substrates are inherently deficient in free sugars, which limits LAB proliferation and acid production. Therefore, supplementation with fermentable carbohydrate sources is essential for effective biological silage production (Lukic et al., 2019; Koirala et al., 2025). Sugarcane molasses, an inexpensive and readily available by-product of the sugar industry, is rich in sucrose and reducing sugars and has been widely used to promote rapid acidification and microbial stability during fermentation, particularly under tropical conditions (Luo et al., 2021; Castro-Montoya et al., 2023). Previous studies have demonstrated that the addition of molasses in combination with LAB can improve fermentation kinetics, silage palatability, and nutritional quality (Vidotti et al., 2002).

Despite these advantages, the quality and stability of fermented fish silage are strongly influenced by the interaction and ratio between the microbial inoculant and the carbohydrate source. Suboptimal combinations may lead to incomplete fermentation, microbial spoilage, or nutritionally inferior products (Zahar et al., 2002). Furthermore, while the nutritional potential of fermented fish silage has been widely reported, its economic feasibility remains a key determinant for adoption by small-scale and commercial aquaculture producers.

Therefore, the present study aimed to optimize the combination of sugarcane molasses and Lactobacillus plantarum for the production of fermented fish silage from fish processing waste. In addition, an economic analysis was conducted to evaluate the cost-effectiveness of the optimized silage production process, with the objective of assessing its potential as a sustainable and economically viable alternative protein source for aquafeed formulation.

Source and preparation of raw materials

Pangas fish processing residuals (15 kg; 5 kg per set), consisting exclusively of viscera, were collected from the Fish Processing Unit, Experimental-cum-Research Farm, College of Fisheries, Guru Angad Dev Veterinary and Animal Sciences University (GADVASU), Ludhiana, and local fish markets. No head, muscle, or other tissue fractions were included in the study. The viscera were specifically selected due to their high content of endogenous digestive enzymes, which facilitate autolysis and enhance the fermentation process. Additionally, fish viscera remain an underutilized and economically less-explored by-product of the fish processing industry, justifying their selection for value addition in the present study. The selection was further supported by preliminary proximate and physicochemical analyses carried out in this study. However, the detailed biochemical composition of the viscera (e.g., proportions of different organs) was not individually characterized and is acknowledged as a limitation. The waste was transported to the laboratory under refrigerated conditions, thoroughly washed with potable water to remove adhering blood and debris, mechanically minced to obtain a homogeneous substrate, and stored under refrigeration until further use.

Proximate, physicochemical, and microbiological analyses were conducted on raw Pangasianodon hypophthalmus muscle tissue and sugarcane molasses prior to fermentation. Proximate composition analyses were performed in accordance with standard AOAC (2022) procedures. Moisture content was determined by oven drying to constant weight, crude protein by the Kjeldahl method using a nitrogen-to-protein conversion factor of 6.25, crude fat by Soxhlet extraction, and ash content by incineration in a muffle furnace at 550 °C for 6 h. Total carbohydrate content was calculated by difference as 100 − (moisture + protein + fat + ash).

Physicochemical parameters, including pH and water activity, were measured for both substrates. The pH was recorded using a calibrated digital pH meter, while water activity was determined at 25 °C using a water activity meter. Microbiological quality was assessed for the raw fish substrate by total plate count, and results were expressed as log CFU g -1.

Experimental design

The experiment followed a two-factor factorial design combining six molasses levels (0, 5, 10, 15, 20, and 25% w/w; M1–M6) with four Lactobacillus plantarum (L.P) inoculation levels (0, 1, 3, and 5% v/w; L1–L4), resulting in 24 treatment combinations as depicted in Figure 1. Treatments without L.P served as controls to evaluate spontaneous fermentation, while increasing inoculation levels enabled assessment of dose-dependent effects of L.P and its interaction with molasses on silage quality.

Preparation and revival of Lactobacillus plantarum

The L. plantarum stock culture, maintained on glucose agar at 4 °C, was aseptically revived in de Man, Rogosa, and Sharpe (MRS) broth autoclaved at 121 °C for 15 min. A loopful of culture was inoculated into 10 mL sterile MRS broth and incubated at 37 °C for 24 h, followed by sub-culturing (1 mL into 10 mL fresh MRS broth) under identical conditions to enhance viability. Actively growing cultures (10 5 –10 7 CFU mL -1) were used as inoculum for silage preparation.

Preparation of biological fish silage

Biological fish silage was prepared in 24 sterile 500 mL conical flasks, each containing 200 g of minced fish waste (Table 1). Molasses, obtained from the Aquaculture Unit, College of Fisheries, was added as a fermentable carbohydrate source at levels ranging from 0 to 25% (w/w). L. plantarum inoculum was applied at 0, 1, 3, and 5% (v/w). The contents were thoroughly mixed to ensure uniform distribution of molasses and inoculum, and water was added as required to maintain suitable moisture content. The mixtures were transferred to airtight containers to establish anaerobic conditions.

Incubation conditions and temperature monitoring

Fermentation was carried out in a BOD incubator at 35–37 ± 1 °C for 30 days, corresponding to the optimal growth range of L. plantarum. Temperature was monitored weekly using a calibrated digital thermometer (± 0.1 °C accuracy), with the probe inserted approximately 5 cm into the core of the silage mass under aseptic conditions. Measurements were recorded in triplicate and averaged.

Monitoring of fermentation

Fermentation progress was assessed at predetermined intervals. Temperature, pH, and water activity were measured weekly; proximate composition, free fatty acids (FFA), peroxide value (PV), and color parameters were analyzed fortnightly; and total plate count (TPC) was determined monthly. Successful fermentation was indicated by stable low pH, uniform texture, characteristic sour aroma, and absence of spoilage. Upon completion, silage samples were stored in sealed containers until analysis.

Analytical procedures

pH measurement

The pH was measured weekly by homogenizing 10 g of silage with 90 mL distilled water (1:10 w/v), allowing equilibration for 10 min, and recording pH using a calibrated digital pH meter (buffers pH 4.0 and 7.0). The electrode was rinsed with distilled water and disinfected with 70% ethanol between measurements.

Water activity

Water activity (aw) was measured weekly using a calibrated water activity meter. Approximately 3–5 g of sample was placed in the sample holder, sealed, and readings were recorded after equilibration. Measurements were performed in triplicate and averaged.

Color measurement

Color attributes (L, a, b*) were measured using a Lovibond Tintometer (Konica Minolta, Japan) following manufacturer calibration guidelines. Measurements were taken at three different locations on each sample, and mean values were recorded.

Proximate composition

Proximate composition was determined following AOAC (2022) standard procedures. Moisture content was estimated by oven drying at 100 ± 2 °C to constant weight. Crude protein was analyzed using the Kjeldahl method (N × 6.25). Crude fat was determined by Soxhlet extraction using petroleum ether (60–80 °C). Ash content was measured by incineration in a muffle furnace at 550 °C. Carbohydrate content was calculated by difference.

Free fatty acids

FFA content (% oleic acid) was determined according to Koniecko (1979). Lipids extracted with chloroform were titrated against 0.1 N alcoholic KOH using phenolphthalein as indicator, and FFA values were calculated accordingly.

Peroxide value

PV was assessed following Koniecko (1979) with minor modifications. Lipid extracts were reacted with potassium iodide in acetic acid–chloroform medium, and liberated iodine was titrated with 0.02 N sodium thiosulfate using starch as indicator. Results were expressed as meq O 2 kg -1 fat.

Microbiological analysis

Total plate count

Total Plate Count (TPC) was determined monthly using the pour plate method with Plate Count Agar (PCA). Ten grams of silage were homogenized in 90 mL sterile saline (0.85% NaCl), followed by serial decimal dilutions. Aliquots (1 mL) were plated, incubated at 35 ± 2 °C for 24–48 h, and colonies (30–300) were enumerated. Results were expressed as log 10 CFU g -1.

Detection of E. coli and Salmonella

Detection of Escherichia coli and Salmonella spp. in fermented fish silage samples was performed using standard culture-based methods (Ayalew et al., 2024). For E. coli, 10 g of sample was homogenized in 90 mL sterile buffered peptone water (BPW) and serially diluted. Aliquots were plated on MacConkey Agar and Eosin Methylene Blue (EMB) agar and incubated at 37 °C for 24 h. Colonies showing characteristic pink colonies and metallic green sheen were considered presumptive E. coli.

For Salmonella detection, 10 g of sample was pre-enriched in 90 mL BPW and incubated at 37 °C for 18–24 h, followed by selective enrichment in Rappaport–Vassiliadis (RV) broth at 42 °C for 24 h. Enriched samples were streaked onto Bismuth Sulphite Agar (BSA) and Xylose Lysine Deoxycholate (XLD) agar and incubated at 37 °C for 24–48 h. Typical colonies (red with black centers) were considered presumptive Salmonella.

All analyses were performed in triplicate, and results were expressed qualitatively as presence/absence per gram of sample. The detection limit of the method was approximately 10 CFU g -1.

Statistical analysis

All experiments were conducted in triplicate (n=3, biological replicates), and results are expressed as mean ± standard deviation (SD). Data were analyzed using a factorial analysis of variance (two-way ANOVA) to evaluate the effects of molasses concentration, Lactobacillus plantarum inoculation level, and their interaction on fermentation, physicochemical, and microbiological parameters. When significant differences were detected (p < 0.05), means were separated using Tukey’s honestly significant difference (HSD) post hoc test. Statistical analyses were performed using appropriate statistical software, and graphical representations were generated to illustrate treatment effects. Graphical analyses were performed using RStudio software.

Yield of fermented fish silage

The yield of fermented fish silage under the optimized formulation (15% molasses + 5% L. plantarum, M4L4) was determined. From 1 kg of whole Pangas fish, 0.42 kg of viscera and internal organs were obtained as processing waste. Fermentation of viscera produced 0.89 kg of liquid silage per kilogram of raw viscera, resulting in an overall yield of 0.37 kg of fermented silage per kilogram of whole fish.

Substrate characterization

A representative subsample of raw Pangasianodon hypophthalmus was analyzed prior to fermentation to establish baseline physicochemical and microbiological characteristics of the substrate. On a wet weight basis, the fish tissue contained 72.5 ± 1.2% moisture, 14.8 ± 0.5% crude protein, 8.3 ± 0.4% crude fat, and 2.8 ± 0.2% ash. Total carbohydrates, calculated by difference, accounted for 1.6 ± 0.3% of the total composition. The initial pH of the substrate was 6.8 ± 0.1, while water activity was recorded as 0.96 ± 0.01 at 25 °C. The total viable microbial load was 5.2 ± 0.3 log CFU g -1, indicating a nutritionally rich and microbiologically active matrix conducive to fermentation.

Food-grade sugarcane molasses was employed as an external carbohydrate source to support fermentation and was procured from a local supplier. The proximate composition of molasses (wet weight basis) comprised 18.0 ± 1.0% moisture, 3.0 ± 0.3% crude protein, negligible crude fat (<0.5%), and 8.0 ± 0.5% ash. Total carbohydrate content, determined using Brix-based measurements, was 62.0 ± 3.0%. The pH of molasses ranged from 4.5 to 5.5, and water activity varied between 0.75 and 0.80 at 25 °C.

Since sugarcane molasses is a rich and economical source of fermentable carbohydrates, and raw pangasius fish serves as an excellent protein source, these substrates were therefore selected and undertaken for the present study.

Temperature changes during fermentation

During the first week, internal silage temperatures ranged from 30.6 °C (M6L3) to 34.2 °C (M1L1) (Figure 2). Temperatures decreased consistently over the 4-week fermentation period. By week 4, maximum and minimum temperatures were 25.3 °C (M1L3) and 20.5 °C (M6L1), respectively. The largest temperature drop of 10.8 °C was observed in M6L1, while M1L1 showed the smallest decline of 9.2 °C.

In the initial fermentation phase (week 1), temperatures remained elevated around 34 °C across all treatments. From weeks 2 to 4, core temperatures gradually decreased, stabilizing between 25–30 °C by the end of the fourth week.

pH changes during fermentation

The pH of all silage treatments decreased progressively over the 4-week fermentation period, reflecting the activity of lactic acid bacteria. Biologically fermented silages supplemented with molasses and Lactobacillus plantarum declined from initial values of 6.5–6.7 to 3.9–4.5 by day 5. Chemically acidified silages reached pH below 4.0 within 24–48 h (Figure 3).

Among the treatments, M5L4 (20% molasses + 5% L. plantarum) and M4L4 (15% molasses + 5% L. plantarum) exhibited the fastest significant acidification, reaching pH 3.95–4.11 by week 1 and maintaining stable acidic conditions thereafter. Treatments with lower molasses or inoculum levels, such as M1L1 (0% molasses + 0% L. plantarum), showed slower acidification, with pH values remaining above 6.5 during the first week and stabilizing between 6.8–6.9 by week 4.

Weekly measurements indicated consistent trends: M1L1 retained the highest pH throughout (6.61–6.92), whereas the M5 series reached the lowest pH values (3.95–4.14). These data demonstrate treatment-dependent fermentation kinetics, with higher molasses and inoculum levels accelerating lactic acid production and rapid pH reduction.

Water activity (a w) during fermentation

The initial water activity (a w) of all treatments at day 0 was highest in the absence of molasses (M1L1–M1L4), with values of 0.98, reflecting the abundance of free water in the silage mass prior to fermentation (Figure 4). Over the 30-day fermentation period, aw decreased in all treatments due to microbial conversion of sugars into organic acids and the osmotic effects of molasses. The lowest a w values (0.84) were recorded in treatments M4L4 (15% molasses + 5% L. plantarum), M5L3 (20% molasses + 3% L. plantarum), and M5L4 (20% molasses + 5% L. plantarum) on day 30. Treatments with lower molasses or inoculum levels showed a more gradual decline, with a w remaining above 0.92 in M1L1–M1L2, reflecting reduced water binding and slower fermentation activity.

Color analysis of silage (L, a, b*)

Color measurements indicated significant variation among treatments at both 15 and 30 days of fermentation (Figure 5). The lightness (L) was significantly higher in M4L2, recording mean values of 27.07 (15 days) and 25.73 (30 days), whereas lower L values were observed in M3L2 and M3L3, with values around 23.07–25.63 (15 days) and 26.00–26.03 (30 days). The red–green component (a) was significantly higher in M4L4, with mean values of 7.93 (15 days) and 8.10 (30 days), indicating increased red/brown pigmentation likely due to intensified microbial activity. The yellow–blue component (b) was significantly higher in M2L1 and M3L4, with values of 17.27 and 14.50 (15 days) and 14.00 and 17.10 (30 days), respectively, reflecting accumulation of yellow-brown pigments. Overall, these results demonstrate that both molasses concentration and L. plantarum inoculum level significantly influenced silage color development over the fermentation period.

Proximate composition of silage

Moisture content

Moisture content of the silage samples varied with treatment and fermentation time. The highest initial moisture was recorded in M1L1 (0% molasses + 0% L. plantarum) at 77.00% on Day 0, while the lowest final moisture was observed in M6L4 (25% molasses + 5% L. plantarum) at 65.43% on Day 30. Across all treatments, moisture levels decreased progressively during fermentation, with higher molasses and inoculum treatments showing greater reductions. Intermediate values were observed in M2–M5 treatments, reflecting gradations based on molasses and L. plantarum concentration (Figure 6).

Crude protein content

Crude protein (CP) content fluctuated over the fermentation period. Initial CP on Day 0 ranged from 15.00% in M1L1 to 16.00% in M6L4 (Figure 7). A temporary decline in CP was observed in treatments with low molasses and inoculum, with the lowest recorded value of 9.23% in M1L2 (0% molasses + 1% L. plantarum) by Day 15. By Day 30, CP levels increased in most treatments, with the highest final protein content of 19.43% recorded in M3L4 (10% molasses + 5% L. plantarum). Intermediate CP values were observed in M2L3–M5L4 treatments, demonstrating variation based on molasses and inoculum levels.

Crude fat content

Crude fat content varied with treatment and fermentation time. Initial fat percentages on Day 0 ranged from 3.00% in M6L1 (25% molasses + 0% L. plantarum) to 4.10% in M1L4 (0% molasses + 5% L. plantarum) (Figure 8). By Day 15, fat content in most treatments reached a minimum, such as 3.01% in M4L1. Subsequently, fat levels increased and stabilized by Day 30, with the highest final value of 4.43% recorded in M6L4 (25% molasses + 5% L. plantarum). All other treatments showed intermediate final fat values ranging from 3.45% to 4.20%, reflecting progressive concentration of residual fat during fermentation.

Ash content

Ash content exhibited treatment- and time-dependent changes. Initial ash values on Day 0 ranged from 2.30% in M3L2 (10% molasses + 1% L. plantarum) to 7.07% in M3L4 (10% molasses + 5% L. plantarum) (Figure 9). A marked decline occurred by Day 15 in most treatments, with the lowest value of 7.43% observed in M4L2 (15% molasses + 1% L. plantarum). By Day 30, ash content stabilized or increased slightly across treatments, with the highest final value of 17.27% recorded in M2L1 (5% molasses + 0% L. plantarum). Intermediate final values ranged between 9.50% and 15.80% depending on the treatment combination.

Carbohydrate content

Carbohydrate content, determined by difference, varied according to treatment and fermentation time. Initial values on Day 0 ranged from 0.50% in M1L1 (0% molasses + 0% L. plantarum) to 15.75% in M6L4 (25% molasses + 5% L. plantarum), reflecting the contribution of molasses as the primary carbohydrate source (Figure 10). By Day 15, carbohydrate percentages increased slightly in most treatments due to initial concentration effects, with the highest peak of 16.04% recorded in M2L1 (5% molasses + 0% L. plantarum). Subsequently, carbohydrate content declined progressively over the fermentation period, reaching minimum final values by Day 30. The lowest final carbohydrate content of 6.27% was observed in M4L4 (15% molasses + 5% L. plantarum), while other treatments showed intermediate final percentages ranging from 7.50% to 10.12%.

Biochemical changes in silage

Free fatty acid value

Free fatty acid (FFA) concentrations varied across treatments and fermentation time. On Day 15, the lowest FFA value of 1.2% was recorded in M6L1 (25% molasses + 0% L. plantarum), while the highest value of 2.4% was observed in M1L4 (0% molasses + 5% L. plantarum) (Figure 11). By Day 30, FFA content generally increased in all treatments, with the highest final FFA of 2.7% recorded in M1L4. Treatments with higher molasses and lower inoculum levels exhibited comparatively lower FFA accumulation. Overall, FFA concentrations increased from Day 15 to Day 30, reflecting progressive lipid hydrolysis during fermentation.

Peroxide value

Peroxide values (PV) were initially low across all treatments, ranging from 3.10 to 4.90 meq/kg on Day 0. By Day 15, PV increased in all samples, reaching a maximum of 4.76 meq/kg in M6L4 (25% molasses + 5% L. plantarum), representing the peak of primary lipid oxidation (Figure 12). By Day 30, PV generally stabilized or slightly increased, with the highest final PV of 5.00 meq/kg still observed in M6L4. Other treatments demonstrated final PV values between 3.90 and 4.85 meq/kg. These results indicate that peroxide formation occurred primarily during the active fermentation phase and reached relative stability toward the end of the 30-day period.

Microbiological quality of silage

Total plate count

At the end of 30 days of fermentation, the total plate count (TPC) of fermented fish silage varied among the 24 treatment combinations of molasses (M1–M6) and Lactobacillus plantarum levels (L1–L4), with values ranging from 6.20 to 7.67 log CFU/g (Figure 13).

Among the treatments without L.P. inoculation (L1), TPC values were moderate, indicating natural fermentation supported mainly by molasses. The counts in these treatments ranged approximately between 6.53 and 7.10 log CFU/g, with slightly higher values observed at increased molasses inclusion levels such as M4L1 (15% M + 0% L.P.) and M6L1 (25% M + 0% L.P.), suggesting that higher carbohydrate availability supported indigenous microbial growth.

In treatments containing 1% L.P. (L2), TPC values generally increased compared to L1. The highest count (7.67 log CFU/g) was recorded in M2L2 (5% M + 1% L.P.), followed by elevated values in M4L2 (15% M + 1% L.P.) and M3L2 (10% M + 1% L.P.), indicating active proliferation of inoculated lactic acid bacteria under moderate molasses supplementation.

For the 3% L.P. treatments (L3), TPC values ranged between 6.43 and 7.47 log CFU/g. Notably, M4L3 (15% M + 3% L.P.) recorded a high value of 7.47 log CFU/g, while lower counts were observed in treatments with either very low or very high molasses levels, such as M1L3 (0% M + 3% L.P.) and M5L3 (20% M + 3% L.P.), suggesting that optimal carbohydrate concentration plays a role in sustaining LAB growth.

In the 5% L.P. group (L4), TPC values varied from 6.20 to 7.50 log CFU/g. The lowest count (6.20 log CFU/g) was observed in M1L4 (0% M + 5% L.P.), indicating that absence of molasses limited microbial proliferation despite higher inoculum levels. Conversely, higher counts were observed in M4L4 (15% M + 5% L.P.) and M6L4 (25% M + 5% L.P.), reflecting sustained fermentative activity under adequate substrate availability.

Overall, treatments with moderate molasses levels (10–15%) combined with 1–3% L.P. generally exhibited higher TPC values, whereas treatments lacking molasses or containing excessive inoculum without sufficient substrate showed comparatively lower counts. These findings demonstrate that both molasses concentration and Lactobacillus plantarum level significantly influenced microbial dynamics and final TPC after 30 days of fermentation.

Detection of E. coli and Salmonella

Neither Escherichia coli nor Salmonella spp. were detected in any of the fermented fish silage samples across all treatment combinations after 30 days of fermentation.

Economic analysis of fish silage production

Cost analysis at laboratory scale confirmed the economic feasibility of the process, with a production cost of ₹93.70 per kg of finished silage when a commercial starter L. plantarum (5% v/w) was used (Table 2). Raw materials (fish waste, molasses and starter culture) constituted the majority of cost (60% of total cost, ₹56.00/kg). Operational costs (electricity, consumables and amortized equipment) accounted for the remaining 40% (₹37.70/kg). The principal cost drivers were the inoculum (₹25.00/kg) and consumables (₹25.00/kg).

The present study demonstrates that Pangas (Pangasianodon hypophthalmus) processing waste can be converted into fish silage via molasses-assisted Lactobacillus plantarum fermentation, with measurable changes in pH, composition, and biochemical indices. These changes indicate that fermentation performance is closely linked to fermentable carbohydrate availability and LAB inoculum level. Although total plate counts remained relatively high (6–7 log CFU g -1), such values are typical of actively fermenting systems and likely reflect LAB predominance rather than contamination. Therefore, microbial counts should be interpreted as indicators of fermentation activity rather than direct measures of microbiological safety.

Fermentation dynamics, temperature regime, and acidification behavior

Rapid acidification is central to biological silage preservation, primarily due to its inhibitory effect on spoilage microbiota. In this study, incubation at 35 ± 2 °C likely enhanced L. plantarum metabolic activity, facilitating early acid production and a quicker transition toward LAB-dominated conditions. This temperature corresponds to the organism’s optimal growth range and may have improved inoculum viability prior to fermentation.

Following inoculation, fermentation under ambient conditions better reflects practical application scenarios. While such conditions are less controlled, the observed pH decline in molasses-supplemented treatments suggests that substrate availability compensated for temperature variability. Similar approaches have been reported by Palkar et al. (2017) and Tazim et al. (2021), where controlled inoculum preparation was combined with ambient fermentation.

The maintenance of low pH under non-incubated conditions indicates effective fermentation; however, it should be interpreted as conditional on adequate carbohydrate supply. These findings support the feasibility of the process but do not eliminate the potential influence of environmental variability in large-scale systems.

Water activity reduction and shelf-stability enhancement

The reduction in water activity (a v) during fermentation likely contributed to microbial stability in conjunction with acidification. Since microbial proliferation depends on available water, the accumulation of solutes and organic acids may have limited the growth of competing microorganisms. Additionally, protein hydrolysis and peptide formation could have influenced water binding within the matrix, although this was not directly quantified.

Carbohydrate source plays a decisive role in modulating moisture dynamics and a v. The role of molasses appears multifaceted, contributing both fermentable sugars and influencing moisture dynamics. Santana et al. (2023) and Palkar et al. (2017) similarly reported that molasses inclusion affects moisture content and fermentation efficiency. Santana et al. (2023) demonstrated significant treatment-dependent variation in moisture, with molasses-based silage exhibiting lower moisture relative to cereal-based substrates. The hygroscopic and viscous properties of molasses likely reduce free water availability while simultaneously serving as a fermentable substrate for lactic acid bacteria. Consistently, Palkar et al. (2017) reported that 15% molasses inclusion stabilized moisture conditions and supported efficient lactic fermentation, underscoring its dual role in substrate provision and water activity regulation. However, the extent to which molasses directly reduces a v versus indirectly influencing it through fermentation remains difficult to separate.

Excessive moisture can dilute acid concentration, delay pH decline, and compromise preservation efficiency (Toppe and Olsen, 2024). Maintaining moisture balance is therefore critical to suppress undesirable microbial proliferation (Andriani et al., 2024), particularly in viscera-based silage systems prone to phase stratification (Santana et al., 2023).

Unlike drying-based preservation (Majumdar et al., 2014), the present study suggests that fermentation alone can enhance stability, although equivalence to dehydration-based methods should be interpreted cautiously without direct comparative data.

Yield improvement and physical characteristics

The observed increase in silage yield likely reflects enhanced solubilization of fish tissues through enzymatic activity. Proteolysis and lipolysis contribute to liquefaction, increasing recoverable material. Similar trends were reported by Raj et al. (2018).

Changes in color during fermentation are consistent with oxidative and non-enzymatic reactions involving pigments and sugars, as described by Tazim et al. (2021). However, these changes are primarily indicative of biochemical transformation rather than direct measures of quality.

Nutritional composition and carbohydrate utilization

The observed increase in crude protein content on a dry matter basis can be attributed primarily to a concentration effect resulting from moisture reduction during fermentation, along with concurrent biochemical transformations. Proteolytic activity mediated by Lactobacillus plantarum facilitates the hydrolysis of complex native proteins into smaller peptides and free amino acids, which may enhance digestibility and biological availability. However, the extent to which hydrolysis alone contributes to net protein enrichment cannot be clearly separated from concentration effects in the absence of nitrogen balance measurements.

During the early stages of fermentation (up to Day 15), the transient decline in measurable crude protein in certain treatments may reflect intensified proteolytic catabolism, particularly under sugar-limited conditions where microbial populations utilize available nitrogenous compounds for metabolic adaptation. This suggests a shift in microbial metabolism toward nitrogen utilization when readily fermentable carbohydrates are insufficient. By Day 30, treatments containing balanced molasses supplementation and higher L. plantarum inoculum exhibited increased crude protein values, indicating improved fermentation efficiency and substrate utilization.

While the contribution of microbial biomass to protein enrichment is plausible, it was not directly quantified in the present study and should therefore be interpreted cautiously. The accumulation of microbial biomass, typically comprising approximately 50–60% protein, may contribute to the observed increase; however, its relative contribution compared to moisture loss and protein transformation remains uncertain. Overall, the data suggest a combined effect of concentration-driven changes and biochemical modification of protein fractions, rather than a single dominant mechanism.

Carbohydrate dynamics further highlight the functional role of molasses as the principal fermentable substrate. The higher initial carbohydrate content in molasses-supplemented treatments confirms its role as a readily available source of sucrose, glucose, and fructose. The apparent increase in carbohydrate percentage during early fermentation (Day 15) is more likely attributable to a concentration effect associated with moisture reduction and loss of volatile components, rather than de novo synthesis of carbohydrates.

In contrast, the pronounced decline in carbohydrate content by Day 30 reflects active microbial catabolism, wherein L. plantarum utilizes fermentable sugars for lactic acid production, along with minor quantities of carbon dioxide. This metabolic conversion is directly associated with acidification and contributes to the stabilization of the silage matrix. However, the relationship between carbohydrate depletion and fermentation success should be interpreted as indirect, as it depends on multiple interacting factors including microbial activity, substrate availability, and environmental conditions.

Greater carbohydrate depletion observed in treatments with higher L. plantarum inoculum suggests enhanced metabolic activity and more efficient sugar utilization. This supports the interpretation of accelerated acidogenesis under optimized conditions. These observations are consistent with Tazim et al. (2021) and Majumdar et al. (2014), who reported that carbohydrate availability influences fermentation efficiency, microbial growth, and acid production.

Collectively, the interaction between carbohydrate supplementation and lactic acid bacterial activity appears to govern the progression of fermentation. However, its influence on final nutritional quality, including protein enrichment and preservation stability, should be viewed as the outcome of multiple interdependent processes rather than a direct causal relationship.

Biochemical changes and lipid stability

Free fatty acid (FFA) and peroxide value (PV) trends provide indirect indicators of lipid hydrolysis and oxidative changes during fermentation. The gradual increase in FFA suggests ongoing lipolytic activity, likely mediated by endogenous enzymes and microbial processes. However, this increase should not be interpreted solely as controlled lipolysis, as FFA accumulation reflects the net balance between lipid hydrolysis and subsequent fatty acid utilization or degradation. Similar trends were reported by Raj et al. (2018), who attributed FFA increases to enzymatic triglyceride breakdown under acidic conditions.

The absence of a marked increase in oxidative indicators suggests limited progression of lipid oxidation under the studied conditions. This observation is consistent with Palkar et al. (2017), who reported no detectable rancidity in molasses-based silage during storage, and Santana et al. (2023), who observed relative lipid stability in low-pH silage systems. However, the interpretation of oxidative stability based solely on PV remains limited, as PV reflects only primary oxidation products and does not account for secondary oxidation compounds.

The mid-fermentation peak in PV may indicate a transient phase of oxidative activity prior to the establishment of more stable, low-oxygen conditions. The subsequent decline in PV could result from the decomposition of primary peroxides into secondary oxidation products rather than a complete cessation of oxidative processes. Additionally, while LAB metabolism and molasses-derived components may contribute to antioxidative effects, their specific roles were not directly evaluated in this study and therefore cannot be conclusively established.

Overall, the observed lipid changes suggest that fermentation conditions may limit extensive oxidative deterioration; however, a comprehensive assessment of lipid stability would require analysis of both primary and secondary oxidation products.

Microbiological quality

The reduction in total plate count in optimized treatments suggests suppression of competing microflora, likely due to acidification and LAB activity. Similar reductions have been reported by Palkar et al. (2017) and Ghosh et al. (2024). Palkar et al. (2017) documented a similar decline in TPC from 10 6 to 10³ cfu/g during biological silage storage, supporting the role of fermentation-driven inhibition of competing microflora. Likewise, Ghosh et al. (2024) reported that Lactobacillus plantarum and related LAB contribute to microbial stability through organic acid and bacteriocin production.

The observed microbial trends are consistent with a shift toward LAB dominance, as also described by Sanni and Adesulu (2013). Although Sanni and Adesulu (2013) observed an initial increase in microbial counts during cereal fermentation, the subsequent dominance of beneficial fermentative microbiota led to a more stable system, consistent with the microbial succession pattern observed in the present study.

In addition to total viable counts, pathogen-specific analysis revealed the absence of detectable E. coli and Salmonella spp. in all fermented silage samples. This observation can be attributed to the combined effects of rapid acidification (pH < 4.5), reduced water activity, and competitive dominance of lactic acid bacteria, which are known to inhibit enteric pathogens through organic acid production and bacteriocin activity. However, given the reliance on culture-based methods, these results should be interpreted within the limits of detection sensitivity. The possibility of sub-detectable populations cannot be excluded, and molecular confirmation would strengthen these findings.

Overall, the results support improved microbial stability but should not be interpreted as absolute pathogen elimination.

Implications for sustainable waste valorization

The findings indicate that biological silage production can convert fish processing waste into a nutritionally improved product under defined fermentation conditions. The integration of biochemical changes, microbial dynamics, and process feasibility supports its potential application within waste valorization frameworks.

However, scalability, consistency under variable environmental conditions, and long-term safety require further validation. Thus, while the process aligns with circular bioeconomy principles, its broader application should be considered in the context of operational and environmental constraints.

Conclusion

This study demonstrated that biological fermentation using molasses and Lactobacillus plantarum is an effective strategy for the valorization of Pangas (Pangasianodon hypophthalmus) fish processing waste into nutritionally rich and microbiologically stable fish silage. The 6 × 4 factorial design clearly established that both molasses concentration and L. plantarum inoculum level significantly influenced fermentation kinetics, biochemical stability, microbial quality, and nutrient retention.

Across all evaluated parameters—pH, water activity, peroxide value, free fatty acids, crude protein, crude fat, ash content, carbohydrate utilization, and total plate count—treatments containing both fermentable carbohydrate. Rapid acidification, sustained low pH (<4.5), reduced water activity, and controlled lipid oxidation collectively ensured silage stability during storage.

Among the 24 treatment combinations, M4L4 (15% w/w molasses + 5% v/w L. plantarum) emerged as the optimal formulation, achieving the most favorable balance between fermentation efficiency, nutritional quality, and microbial stability. This treatment exhibited rapid and stable acidification, the highest crude protein retention, minimal lipid degradation (low FFA and PV), acceptable water activity, and controlled microbial counts. The interaction effects observed between molasses and L. plantarum primarily influenced the rate of fermentation rather than the final stabilized state, confirming that sufficient fermentable substrate combined with a robust L.P. inoculum is critical for efficient silage production.

Economic evaluation further confirmed the feasibility of biological fish silage production. Pure L. plantarum cultures offer superior process control and reproducibility for lab-scale applications. The optimized formulation demonstrated favorable production costs at laboratory scales, reinforcing its suitability for sustainable waste utilization and feed production.

Overall, the findings validate biological fish silage production as a viable, low-cost, and environmentally sustainable approach for managing fish processing waste while generating value-added aquafeed ingredients.

Future prospects

The outcomes of this study open several important avenues for future research and application. Foremost, in vivo feeding trials are required to evaluate the nutritional efficacy of the optimized silage (M4L4) as a partial or complete replacement for conventional fishmeal in aquaculture diets. Growth performance, feed conversion ratio, nutrient digestibility, immune response, and health status of cultured species such as Pangas should be systematically assessed.

Further research should also focus on microbial community dynamics during fermentation using advanced molecular tools such as 16S rRNA sequencing and metabolomic profiling. These approaches would provide deeper insights into microbial succession, functional pathways, and bioactive metabolite production responsible for silage stability and nutritional enhancement.

Long-term shelf-life studies under varied storage conditions (ambient, refrigerated, and extended storage durations) are necessary to confirm product stability beyond 30 days. Additionally, exploration of alternative low-cost carbohydrate sources (e.g., agro-industrial by-products) could further improve economic viability without compromising fermentation efficiency.

From an applied perspective, pilot-scale validation and techno-economic analysis are essential to support large-scale adoption. Integration of this technology into decentralized fish processing units and rural waste management systems could significantly reduce environmental pollution while enhancing resource efficiency. A comprehensive life cycle assessment (LCA) would further quantify environmental benefits and support policy-level adoption.

In conclusion, biologically fermented fish silage represents a promising pathway for circular bioeconomy development in the fisheries sector, combining waste valorization, feed security, and environmental sustainability.