Liver and Muscle Transcriptomes, From: NAD repletion boosts dietary restriction to improve stem cell function and lifespan in aged mice

This is the R code used to generate the liver and muscle transcriptomes for this paper - the raw counts and differentially expressed gene tables. Raw transcriptome (fastq) data are here: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE278311 All code was run on R 4.5.0 (April 2025) using the then-current versions of all packages (Rsubread: 2.23.3; dplyr: 1.1.4). Code can be run on a regular desktop computer, however at least 24 GB of RAM are necessary. A single alignment will take around 10 minutes on a normal 2025-era computer (e.g. base Macbook Pro) if the minimum RAM conditions are met. No installations are necessary other than R. The full output of the code is included ( 02a_Jena_data_liver_WholeGenomeMouse_RNAseq_Counts.csv and 02b_Jena_Muscle_RNAseq_WholeCounts_Counts.csv ) but theoretically a user may want to re-run the code from scratch if using a newer version of the mouse genome assembly (mm10 was used) or gene mapping database (release 103 was used). However, the genome assemblies and coding gene maps have been relatively stable for more than 10 years at this point, so we suggest that re-running the entire alignment is more of a homework exercise than something that is likely for a reader to need to perform. To run the code, there are two steps: First, follow instructions in the 01_FASTQ_to_Counts_File.r in R. This requires having PREVIOUSLY downloaded all raw FASTQ sequence files from GEO ( https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE278311 ). Once this is done (roughly 10 minutes per sample - depending on computer and again minimum 24 GB of RAM), readers will generate one intermediate "BAM" files for each sample. This step can be done in parallel if necessary - although typically we just run in series on a computer overnight for simplicity. Next, and still within the 01_FASTQ_to_Counts_File.r file, the featureCounts function will be run on all samples together, which takes about 2 minutes per sample on a normal desktop computer and CANNOT be done in parallel. These steps can be massively sped up if necessary by e.g. using a compute cluster. Then, the output CSV file is simply a raw data matrix containing a list of all genes from the GTF file, and how many times each gene was observed (counted) in each sample. This feature counts file is then used as input for step 02 - step 02a and 02b are IDENTICAL, except that one is handling the liver data and one is handling the muscle data. All steps in 02a_DESeq2_BesmaCode_LiverData.R and 02b_DESeq2_BesmaCode_Muscle.R run essentially instantaneously on any modern computer, and there are no meaningful requirements in terms of RAM or processing power. The only code that may need to be changed is the group assignments, should readers be interested in comparing groups that we did not focus on in this study (e.g. comparing old dietary restricted B6 mice to young ad libitum diet HET3 mice). The lines where the code would need small modifications are indicated within the R scripts for 02a and 02b.