Page 73 - Read Online

Monaco et al. J Environ Expo Assess 2024;3:18 https://dx.doi.org/10.20517/jeea.2024.10 Page 5 of 18

manufacturer’s recommendations (PacBio, Menlo Park, CA). The pooled, individually barcoded amplicons
were converted to a library with the SMRTBell Express Template Prep kit 2.0 (PacBio). The library was
quantitated with Qubit 3.0 Fluorometer and run on a Fragment Analyzer to confirm the presence of DNA
fragments of the expected size. The library was sequenced on an SMRTcell 8M in a PacBio Sequel II with
10 h movie time. The circular consensus analysis was performed from the bam file with SMRTLink V8.0
using the following parameters: ccs --min-length 1200 --max-length 2000 --min-passes 3 --min-rq 0.999.
Demultiplexing was done with lima using default parameters (PacBio).

Sequence processing
Sequences were processed using DADA2 R package and QIIME 2 pipeline [29,30] . Demultiplexed reads
obtained from the sequencing facility were subjected to primer removal, dereplication, denoising, and
[29]
chimera checking using DADA2 package of R (version 1.14.1) . The amplicon sequence variant (ASV)
table and representative sequences generated from DADA2 were imported and further analyzed in QIIME
[30]
pipeline (version 2020.6) . The representative sequences were aligned, and a phylogenetic tree was
constructed from filtered alignment as previously . Alpha diversity and UniFrac distance metrics were
[31]
computed through the plugin diversity with the core metrics-phylogenetic method on the ASV table and
phylogenetic tree. Samples were rarefied to an equal number of reads (10,956) when α-diversity and UniFrac
distance metrics were calculated. Taxonomic assignments were performed using plugin feature-classifier
with classify-sklearn method [32,33] , in which a pre-trained Naïve-Bayes classifier with SILVA 138 99%
operational taxonomic units (OTUs) full-length sequences was used (https://www.arb-silva.de/) .
[34]
Taxonomic bar plots were generated using the plugin taxa with the taxa barplot method. The bar plots were
visualized, and the count tables at the phylum and genus levels were downloaded using QIMME 2 View
(https://view.qiime2.org/) for differential abundance analysis.

Microbiota statistical analyses
Differences in overall bacterial communities among treatment groups (beta-diversity) were evaluated with
principal coordinate analysis (PCoA) and permutational multivariate analysis of variance (PERMANOVA).
PERMANOVA was performed on weighted UniFrac distances in QIIME 2 using plugin diversity with beta-
[30]
group-significance method . Alpha diversity [Shannon’s indices, Evenness, Faith’s phylogenic diversity
(PD)] was analyzed using the PROC MIXED procedure of SAS (version 9.4; SAS Institute). For differential
abundance analysis, counts of bacterial taxa were assessed with pairwise comparisons using a negative
binomial Wald test implemented in the R package DESeq2 . Genera with mean relative abundance < 0.01%
[35]
and present in < 20% of samples were removed from the differential abundance analysis. Data are reported
as means ± SEM or log2 fold change. Statistical significance was defined as P ≤ 0.05 for alpha and beta
diversity analysis. For differential abundance of bacterial taxa, P-values were adjusted (Padj) for multiple
testing by the Benjamini–Hochberg procedure, and a Padj ≤ 0.10 was considered statistically significant.

RESULTS
Growth, formula intake, and organ weights
There were no differences in BW over the course of the study between the piglets dosed with 20 or 200 mg
DEHP/kg BW/day compared to CON [Figure 1]. Total BW gain, calculated as the difference in weights
between PND23 and PND2, did not differ among the groups, and the gains were 6.2 ± 0.07, 6.2 ± 0.1 and 6.0
± 0.1 kg for CON, DEHP20 and DEHP200, respectively. Similarly, the average daily weight gain (ADG) in
three periods of seven days each did not differ among the groups [Supplementary Table 1]. Formula intake
was not affected by oral DEHP (data not shown). DEHP exposure at either dose did not affect intestinal
length or organ weights, normalized to BW, compared to CON [Table 1]. BW and ADG did not differ
between males and females. Thus, all remaining statistics did not include sex as a variable.