Additional file 4 - link for raw sequence data

Microbiota analysis was carried out in R using the DADA2 pipeline (Callahan et al., 2016), and taxonomic categories were assigned using the Silva Database (release 138.1) as a reference for the assignment (Quast et al., 2012). Alpha (Shannon, Chao1, and Simpson indices) and Beta diversity (calculated as Bray Curtis distance matrix), as well as the abundance of taxonomic categories, were calculated and analysed with R software 3.6, using the PhyloSeq (McMurdie et al., 2013), Vegan (Dixon, 2003), and microbiomeMarker (Cao et al., 2022) packages. Differences between groups regarding Alpha diversity indices were tested using a linear mixed model that included category (IUGR, NORM), time (T1, T2, T3, T4), and their interaction as fixed factors, with the litter of origin considered as a random factor. Then, a separate model within each time point was performed including only the category as a fixed factor and the litter of origin as a random factor. For Beta diversity, a dissimilarity matrix was constructed using the Bray–Curtis distance matrix as metrics, and the results were plotted using a non-metric multidimensional scaling (NMDS) plot. Furthermore, the betadisper test was conducted to test differences in the samples’ dispersion among the groups. Differences were tested using a PERMANOVA (Adonis) model with 9,999 permutations, including category, time, and their interaction, and litter of origin as factors. A second PERMANOVA model was then performed within each time point, including the category as a factor. For the differential analysis of taxa, the LEfSe algorithm was used at the genus level. Specifically, taxa were considered significant if they possessed a Linear Discriminant Analysis (LDA) score &gt; 4 and a Padj &lt; 0.05 within each time point, including the category as a factor.References: Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP: DADA2: High-resolution sample inference from Illumina amplicon data. Nature Methods 2016, 13:581-583.Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, Peplies J, Glöckner FO: The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Research 2012, 41:D590-D596.McMurdie PJ, Holmes S: phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PloS One 2013, 8:e61217.Dixon P: VEGAN, a package of R functions for community ecology. Journal of Vegetation Science 2003, 14:927-930.Cao Y, Dong Q, Wang D, Zhang P, Liu Y, Niu C: microbiomeMarker: an R/Bioconductor package for microbiome marker identification and visualization. Bioinformatics 2022, 38:4027-4029.<br><br>

本研究依托R语言，采用DADA2流程（DADA2 pipeline）开展微生物组分析（Callahan等，2016），并以Silva数据库（Silva Database，版本138.1）作为分类注释的参考数据集（Quast等，2012）。本研究采用R 3.6软件，结合PhyloSeq（McMurdie等，2013）、Vegan（Dixon，2003）与microbiomeMarker（Cao等，2022）三款R包，计算并分析α多样性（涵盖Shannon指数、Chao1指数与Simpson指数）、β多样性（以Bray-Curtis距离矩阵计算）以及分类单元丰度。

针对α多样性指数的组间差异，本研究采用线性混合模型进行检验：模型以分组类别（宫内生长受限组IUGR、正常对照组NORM）、时间点（T1、T2、T3、T4）及其交互项作为固定因子，以产仔窝作为随机效应因子。随后，在每个时间点内单独构建模型，仅以分组类别作为固定因子、产仔窝作为随机效应因子。

针对β多样性分析，本研究以Bray-Curtis距离矩阵作为度量标准构建样本相异矩阵，并通过非度量多维尺度分析（non-metric multidimensional scaling, NMDS）对分析结果进行可视化；此外，采用betadisper检验分析各组样本离散程度的差异。组间差异通过置换多元方差分析（PERMANOVA，又称Adonis）进行检验，置换次数设定为9999次，模型纳入分组类别、时间点及其交互项，以及产仔窝作为因子；随后在每个时间点内再次构建PERMANOVA模型，仅以分组类别作为因子。

针对分类单元的差异分析，本研究在属水平采用线性判别分析效应大小（Linear Discriminant Analysis Effect Size, LEfSe）算法。具体而言，在每个时间点内，当分类单元的线性判别分析（Linear Discriminant Analysis, LDA）得分>4且校正后P值（Padj）<0.05时，即判定为具有显著性差异，模型纳入分组类别作为因子。

### 参考文献
1. Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP. DADA2：基于Illumina扩增子数据的高分辨率样本推断. 《自然-方法》（*Nature Methods*），2016，13：581-583.
2. Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, Peplies J, Glöckner FO. Silva核糖体RNA基因数据库项目：优化的数据处理与网页工具. 《核酸研究》（*Nucleic Acids Research*），2012，41：D590-D596.
3. McMurdie PJ, Holmes S. phyloseq：用于微生物组普查数据可重复交互式分析与可视化的R包. 《公共科学图书馆·综合》（*PloS One*），2013，8：e61217.
4. Dixon P. Vegan：面向群落生态学的R函数包. 《植物科学杂志》（*Journal of Vegetation Science*），2003，14：927-930.
5. Cao Y, Dong Q, Wang D, Zhang P, Liu Y, Niu C. microbiomeMarker：用于微生物组标志物识别与可视化的R/Bioconductor包. 《生物信息学》（*Bioinformatics*），2022，38：4027-4029.