Fastqc

For the RNA-seq data, low-quality (Q < 30) reads and sequencing adapters were removed with Trimmomatic [77 (link)] and the quality of raw reads was assessed with FastQC (version 0.11.5, Babraham Bioinformatics, UK). Clean reads were then mapped to the human reference genome hg19 Refseq using Hisat2 [78 (link)]. FeatureCounts [79 (link)] was used to count reads mapped to individual genes and only uniquely mapped reads were used in the counting step. Genes with zero expression across all samples were omitted from the correlation and differential expression analysis. Fragments per kilobase of exon per million mapped reads (FPKM) was calculated to represent gene expression levels.
Similarly for the RRBS data, quality control of the RRBS data was performed using FastQC (version 0.11.5, Babraham Bioinformatics, UK), and low-quality reads were removed with Trim Galore (version 0.5.0, Babraham Bioinformatics, UK). Then, trimmed reads were aligned and mapped using Bismark [80 (link)] to the human hg19 reference genome. The percentage methylation level was calculated by #C/(#C + #T), where #C is the number of methylated reads and #T is the unmethylated reads. Then, only the CpG sites with a read coverage > 10, a quality score > 20, and appeared at least in 10 samples among each group, using the parameter “min.per.group = 10,” were kept for downstream analysis.

The quality of the raw read data was checked using FastQC (Babraham Bioinformatics; FastQC/">http://www.bioinformatics.babraham.ac.uk/projects/FastQC/, accessed on 20 September 2020). The sequences of adapters and low quality reads were removed using Trimmomatic v.0.38 [26 (link)] to isolate clean reads. The reference genome of Sus scrofa (v.11.1) was used in the FASTA format from Ensemble (http://www.ensembl.org/, accessed on 20 September 2020). Clean, paired-end reads were mapped against an indexed reference genome using HISAT2 [27 (link)], which is a sensitive and fast alignment program for next-generation sequencing reads. The raw counts corresponding to the genes for each library were calculated based on exons and the Sus_scrofa v98 GTF file was taken as the genomic annotation reference file using the feature count function of the R package ‘Subread’ [28 (link)].
Raw counts were normalized using the trimmed means of M values (TMM) method [29 (link)] in the R package ‘edgeR’ [30 (link)], following which DEG profiling was conducted by comparing the normalized read counts between the 10 W and 26 W groups. The false discovery rate (FDR) was calculated using the Benjamin–Hochberg procedure. Significant DEGs were extracted by applying the thresholds of FDR < 0.05 and absolute log₂ fold change (FC) ≥1. The overall expressions of the genes were visualized using the R package ‘ggplot’.

Raw sequencing data were acquired from the North Texas Genome Center at the University of Texas at Arlington and McDermott Sequencing Core at UT Southwestern in the form of binary base call (BCL) files. Raw BCL files were then demultiplexed with 10x Genomics i7 indices (used during library preparation) using Illumina’s bcl2fastq v2.19.1 and “cellranger mkfastq” from 10x Genomics CellRanger v3.0.2 tools. Extracted paired-end reads (28 bp long R1–16 bp 10x cell barcode and 12 bp UMI sequence information, 124 bp long R2—transcript sequence information from cDNA fragment) were first checked for read quality using FastQC v0.11.5 (FastQC, Babraham Bioinformatics, URL: FastQC">https://www.bioinformatics.babraham.ac.uk/projects/FastQC). Extracted paired-end reads were then aligned to the reference human genome (GRCh38.p12) from University of California Santa Cruz (UCSC) genome browser and reference human annotation (Gencode v28) and counted using “cellranger count” from 10x Genomics CellRanger v3.0.2 tools. Since the nuclear transcriptome contained unspliced transcripts, reads mapping to a pre-mRNA reference file were counted. The resulting raw UMI count matrix contains genes as rows and nuclei as columns and was further used for downstream analysis.

Upon RNA integrity checking and quantification (integrity of RNA samples were measured with an Agilent 2100 Bioanalyzer and RIN (RNA Integrity Number) of >7 is required), the total RNA samples were subjected to library construction (insert size ~450 bp, strand-specific using the dUTP protocol; Parkhomchuk et al. 2009 (link)) and Illumina HiSeq sequencing (150 bp paired-end). The quality of the obtained raw sequencing reads was first examined using FastQC (Babraham Bioinformatics; FastQC/">www.bioinformatics.babraham.ac.uk/projects/FastQC/) ver. 0.11.3. Then, a series of quality filtering steps were conducted on the raw reads, including the removal of adapter sequences, removal of reads containing >5 % of N bases, trimming of bases with Phred Q score <20 (i.e. error rate ≥1 %) and removal of sequences with average Phred Q score <20. Only paired-end reads with length ≥50 bp were kept. All filtering steps were conducted using BBDuk embedded in BBTools (jgi.doe.gov/data-and-tools/bbtools) ver. 37.76. The quality-filtered reads were considered high-quality (‘HQ’) and were used for the subsequent assembly and mapping steps.

Read quality was controlled by FASTQC (Babraham Bioinformatics) (FASTQC">http://www.bioinformatics.bbsrc.ac.uk/projects/FASTQC) using default values. Raw reads were filtered for quality (mean phred > 20) and trimmed 10 bp on each end using custom Perl scripts, reducing each read to 80 bp. Paired-reads were then used to estimate the genome size using the program khmerfreq (kmer = 17). The trimmed reads were then assembled using Velvet 1.0.09 [58 (link)] and SoapDenovo [59 (link)] with a range of kmer lengths (57–64) the final assembly being based on assembly size, number of contigs and contig size compared to M. paratuberculosis K10 (Accession number AE016958).

RNA samples were sequenced at Vertis Biotechnologie, where cDNA libraries were made for Illumina sequencing. Single-end 75 bp sequencing was performed using an Illumina NextSeq 500, with approximately 10 million reads obtained per library. The fastq files were run through Trimmomatic to remove adaptors and low-quality reads [38 (link)]. fastqc (Babraham Bioinformatics; http://www.bioinformatics.bbsrc.ac.uk/pro- jects/fastqc/) was run to evaluate the quality of the RNA-seq. Bowtie2 [43 (link)] with default parameters was used to align raw sequence reads to the LacA genome. The alignment was converted to BAM format using SAMtools [44 (link)]. DEseq2 [45 (link)], run in the R environment [40 ], was used to identify differentially expressed genes using a Wald test, followed by a Benjamini and Hochberg procedure. All sets were defined by a false discovery rate of 10 %. Visualization of sequencing reads and calculations of reads per kilobase million (RPKM) were performed in Artemis [46 (link)].