The Zoea-II syndrome samples were obtained from hatcheries located near Marakkanam Tamilnadu, India (12.19° N, 79.92° E). The larvae of zoea stage 1 (50 zoeas per sample) displaying the clinical symptoms as stated above were collected at 2-, 4-, 6- and 8-days post-infection (ZIIS2, ZIIS4, ZIIS6 and ZIIS8). The control group is devoid of ZS infection (ZIISZN). With regard to white feces syndrome, a total of five samples, comprising two diseased (D1 and D2), two recovered (R1 and R2) and one healthy sample (C1) were collected from the shrimp farms located near Nellore of Andhra Pradesh India (14.47° N, 80.09° E). The samples were identified based on the symptoms displayed for each condition, and collected at 45 to 80 days of culture, covering different stages of infection. Pooled gut contents from three biological replicates were collected from each sample, representing all three selected health/disease conditions. All the collected shrimp samples of ZS and WFS groups were stored at -80°C for further analysis.

Samples were subject to CTAB and Phenol: chloroform method followed by RNase A treatment to isolate the total DNA and the quality was checked using NanoDrop 2000 (Thermo Scientific, Brea, CA, United States) followed by 16s rRNA gene amplification. The primers targeting the 16s V3-V4 regions (F -GCCTACGGGNGGCWGCAG; R- ACTACHVGGGTATCTAATCC) designed by Eurofins (Bisht et al., 2018) were used for PCR amplification, and 3 µl of the PCR product was checked using 1.2% agarose gel. This was followed by library preparation using the amplicon library preparation kit Nextera XT Index by Illumina Inc. (San Diego, CA, United States) in accordance with the 16s metagenomic Sequencing Library preparation guide Part #15044223 Rv. B. The amplicon libraries were purified by AMPure XP beads quantified using Qubit Fluorometer and analyzed using the Agilent 4200 TapeStation (Santa Clara, CA, United States) with D1000 Screen tape. The libraries were subject to 2x300 paired ends sequencing using the Illumina MiSeq platform (San Diego, CA, United States).

The quality of the raw reads was assessed using the FastQC v.0.11.8 tool (https://www.bioinformatics.babraham.ac.uk/projects/fastqc ). Further, the raw reads with inadequate quality standards were filtered using Trimmomatic ver. 0.39 (Bolger et al., 2014). The adapters were removed with ILLUMINACLIP option using built-in NextEra PE adapter file, and the low-quality bases were trimmed using the SLIDINGWINDOW option with a phred score threshold of 25 for every 5 bases. Finally, the MINLEN option was used to remove all the sequences that had a length below 50 bases.

The 16s amplicon sequence analyses of the samples were performed using the Quantitative Insights Into Microbial Ecology (QIIME) software, an open source microbiome analysis pipeline (Caporaso et al., 2010). The trimmed high-quality data were subjected to sequence preprocessing such as joining paired ends, converting from fastq to fasta files and combining the sequence with QIIME identifiable label. The pre-processed sequences were further subjected to OTU classification (operational taxonomic units) with open referencing method against the SILVA v132 database using default parameters. Further, taxonomy summarization of the OTU table was carried out followed by alpha and beta diversity analysis (https://github.com/biocore/qiime). The calculated alpha diversity indices include Shannon, Chao1, Simpson’s diversity and evenness, fisher alpha and goods coverage, while the beta diversity indices include the weighted and unweighted unifrac distances ( Supplementary Figures 1 , 2 ).

The function prediction was carried out using the PICRUSt (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States, Langille et al., 2013) tool in three steps. The PICRUSt predictions were based on the OTUs generated using closed referencing method with default parameters ( Supplementary Table 5 ). At first, the OTU table is normalized based on copy number abundance, while the second step involves prediction of Kyoto Encyclopedia of Genes and Genomes (KEGG) (Kanehisa and Goto, 2000) and Cluster of Orthologues Groups (COG) (Tatusov et al., 2003) functional groups. Finally, the categorization of KEGG and COG functions at different levels was carried out (https://github.com/picrust/picrust). The results of QIIME and PICRUSt pipelines are tabulated and visualized using vegan, ggplot2, reshape2, RColorBrewer, and vennDiagram libraries of R (Oksanen, 2007; Wickham, 2007; Chen and Boutros, 2011; Neuwirth and Neuwirth, 2014; Wickham, 2016). Heatmaps on KEGG annotations of PICRUSt are generated using the online tool Morpheus (https://software.broadinstitute.org/morpheus/)

An average of 418 K and 205 K paired-end reads were obtained for ZS and WFS respectively. On trimming, these were reduced to 341 K and 180 K. Finally, high quality reads of 81.16% - 82.31% for ZS and 87.03% – 88.83% for WFS are retained from the raw reads. The Q30 stats for ZS and WFS samples were observed to be in the ranges of 95.65 - 95.71 and 95.19 - 95.44 percentages respectively ( Supplementary Table 1 ).

Amplicon sequence analysis results revealed an average read classifications of 87100 and 190371 for WFS and ZS respectively. The taxonomic profile revealed that all the samples were dominated by Proteobacteria and Firmicutes irrespective of disease state and life stages. In addition, the phyla Bacteroidetes and Actinobacteria were found to be high in ZS samples, while Tenericutes, and Cyanobacteria were high in WFS samples.

Family level abundances revealed a dominance of Planococcaceae, Moraxellaceae, Flavobacteriaceae and Rhodobacteraceae in ZS, while Mycoplasmataceae, Vibrionaceae, and Pirellulaceae were found to be predominant in all WFS samples. In addition, Streptococcaceae and Enterococcaceae were observed to be high in recovered and healthy samples of WFS respectively.

Psychrobacter, Planomicrobium, Pedobacter, and Glutamicibacter are found to be predominant genera of ZS samples, whereas Candidatus Bacilloplasma, Photobacterium, Lactococcus, Enterococcus, and Vibrio were predominant in WFS ( Figure 1 ).

The complete taxonomic profile for both the sample groups are provided in Supplementary Tables 2 and 3 for ZS and WFS respectively. Additionally, sample-wise microbial profiles are depicted in the form of krona charts and are made available in Supplementary Data Sheets 1 and 2 . The diversity analysis results revealed high diversity indices for healthy groups than diseased groups for ZS and WFS samples ( Supplementary Table 4 ). Among the diversity indices calculated for ZS, richness and the fisher’s alpha indices, were high in control sample and gradually declined with the increase in days after infection. However, a sudden rise in diversity was noticed on the 8^th day of infection ( Supplementary Figure 1A ). The richness and fisher alpha indices depicted for WFS samples clearly indicated the higher diversity of healthy and recovered samples compared to diseased samples ( Supplementary Figure 1B ).

In concurrence with earlier studies, genus Candidatus Bacilloplasma and Photobacterium were observed to be high in the diseased samples and Lactococcus were high in recovered sample group of WFS.

Candidatus Bacilloplasma is a lineage of the Mollicutes which are identified as opportunistic pathogens and have earlier been reported to be high in shrimp affected with WFS (Hou et al., 2018). Candidatus Bacilloplasma was first described by Kostanjsek et al. (2007) colonising the cuticular surface of the terrestrial isopod Porcellio scaber (Crustacea: Isopoda), but it had also been found in several marine species. In the Chinese mitten crab Eriocheir sinensis, Candidatus Bacilloplasma was the dominant genus in the midgut (Dong et al., 2018). As this is where food digestion primarily occurs, the presence of Candidatus Bacilloplasma suggests a possible relationship with the function of the digestive tract. In general, Mollicutes represent a common bacterium class in the intestinal tract of arthropods which is beneficial to the host, with some species associated with an increased performance in the adaptive response to food-limiting conditions. Therefore, the increase in abundance of Candidatus Bacilloplasma in diseased crustaceans may represent a compensatory strategy to enhance the absorption of nutrients when they stop feeding.

On the other hand, many species of Vibrio are commensal microorganisms and are involved in the digestion of a wide range of recalcitrant biopolymers, such as chitin, cellulose and alga-derived compounds, as well as lipids and carbohydrates (Gatesoupe et al., 1997). But also, many species of Vibrio are opportunistic pathogens that impact the physiology and health of their hosts.

Similarly, Photobacterium damselae was found to be associated with many marine animal diseases (Terceti et al., 2018). While, Lactococcus was reported to be among the healthy animals’ gut contents which indicates the recovering health status of the shrimps (Piamsomboon and Han, 2022). In the ZS samples, the genus Psychrobacter was consistently high irrespective of the disease stage, in an earlier study Psychrobacter has been reported to have evolved from a pathobiont ancestor (Welter et al., 2021). This genus has been observed in marine mammals’ skin, respiratory, and gut and has also been reported to cause opportunistic infection in mammals (Welter et al., 2021).

The Functional analysis revealed the abundant KEGG categories related to cellular processes, environmental information and processing, metabolism and organismal system in both WFS and ZS samples ( Supplementary Table 6 ). In case of WFS samples, the functional pathways were highly abundant in the recovered samples followed by control samples in comparison to the diseased samples ( Supplementary Table 6 ). Similar observations have been made in earlier studies where the pathways, especially the ones related to metabolism and biological processes, were high in healthy samples and low in disease-affected samples (Wang et al., 2019). Similar trend was observed for ZS with an exception at 8 days post-infection.

KEGG ontology terms (level2) for all the samples of both the diseases are depicted in Figure 2 . Though there is no clear trend observed with regard abundance of functional categories with disease progression, eight days after infection in case of Zoea II syndrome and recovered samples in case of white feces syndrome have reveled higher abundances. The abundant functional categories of previously mentioned disease stages include Carbohydrate Metabolism, Amino Acid Metabolism, Energy Metabolism, Replication and Repair and Membrane Transport. Among the sub-functional classes’ (level 3) transporters, ABC transporters, DNA repair and recombination proteins, purine metabolism and ribosomes are found to be commonly observed in both the diseases with a variation in the abundances. ( Supplementary Figures 3 , 4 ).