This project consisted of a differential expression analysis involving 4 expression features: genes, transcripts, exons, and exon-exon junctions. The main goal of this study was to explore the effects of prenatal smoking and nicotine exposure on the developing brain of mouse. As secondary objectives, this work evaluated the affected genes by each substance in adult brain in order to compare pup and adult results, and the effects of smoking exposure on adult blood and brain to search for overlapping biomarkers in both tissues. Finally, mouse findings were compared against human results from other studies.
Experimental design of the study. A) 21 pregnant mice were split into two experiments: in the first one prenatal nicotine exposure (PNE) was modeled administering nicotine (n=3) or vehicle (n=3) to the dams during gestation, and in the second maternal smoking during pregnancy (MSDP) was modeled exposing dams to cigarette smoke during gestation (n=8) or using them as controls (n=7). A total of 137 pups were born: 19 were born to nicotine-administered mice, 23 to vehicle-administered mice, 46 to smoking-exposed mice, and 49 to smoking control mice. 17 nonpregnant adult females were also nicotine-administered (n=9) or vehicle-administered (n=8) to model adult nicotine exposure, and 9 additional nonpregnant dams were smoking-exposed (n=4) or controls (n=5) to model adult smoking. Frontal cortex samples of all P0 pups (n=137: 42 for PNE and 95 for MSDP) and adults (n=47: 23 for the nicotine experiment and 24 for the smoking experiment) were obtained, as well as blood samples from the smoking-exposed and smoking control adults (n=24), totaling 208 samples. Number of donors and samples are indicated in the figure. B) RNA was extracted from such samples and bulk RNA-seq experiments were performed, obtaining expression counts for genes, transcripts, exons, and exon-exon junctions.
We hope that this repository will be useful for your research. Please use the following BibTeX information to cite this code repository as well as the data released by this project. Thank you!
Molecular impact of nicotine and smoking exposure on the developing and adult mouse brain
Daianna Gonzalez-Padilla, Nicholas J. Eagles, Marisol Cano, Geo Pertea, Andrew E. Jaffe, Kristen R. Maynard, James T. Handa, Keri Martinowich, Leonardo Collado-Torres.
bioRxiv (TODO DOI); doi: https://doi.org/10.1101/TODO
@article {TODO (once we have a pre-print)
}
Summary of analysis steps across gene expression feature levels:
-
1. Data processing: counts of genes, exons, and exon-exon junctions were normalized to CPM and log2-transformed; transcript expression values were only log2-transformed since they were already in TPM. Lowly-expressed features were removed using the indicated functions and samples were separated by tissue and age in order to create subsets of the data for downstream analyses.
-
2. Exploratory Data Analysis (EDA): QC metrics of the samples were examined and used to filter the poor quality ones. Sample level effects were explored through dimensionality reduction methods and segregated samples in PCA plots were removed from the datasets. Gene level effects were evaluated with analyses of variance partition.
-
3. Differential Expression Analysis (DEA): with the relevant variables identified in the previous steps, the DEA was performed at the gene level for nicotine and smoking exposure in adult and pup brain samples, and for smoking exposure in adult blood samples; DEA at the rest of the levels was performed for both exposures in pup brain only. DE signals of the genes in the different conditions, ages, tissues, and species (using human results from
$^1$ Semick et al., 2020) were contrasted, as well as the DE signals of exons and transcripts vs those of their genes. Mean expression of DEGs and non-DEGs genes with and without DE features was also analyzed. Then, all resultant DEGs and DE features (and their genes) were compared by direction of regulation (up or down) between and within exposures (nicotine/smoking); mouse DEGs were also compared against human genes associated with TUD from$^2$ Toikumo et al., 2023. -
4. Functional Enrichment Analysis: GO & KEGG terms significantly enriched in the clusters of DEGs and genes of DE transcripts and exons were obtained.
-
5. DGE visualization: the log2-normalized expression of DEGs was represented in heat maps in order to distinguish the groups of up- and down-regulated genes.
-
6. Novel junction gene annotation: for uncharacterized DE junctions with no annotated gene, their nearest, preceding, and following genes were determined.
Abbreviations: Jxn: junction; Tx(s): transcript(s); CPM: counts per million; TPM: transcripts per million; TMM: Trimmed Mean of M-Values; TMMwsp: TMM with singleton pairing; QC: quality control; PC: principal component; DEA: differential expression analysis; DE: differential expression/differentially expressed; FC: fold-change; FDR: false discovery rate; DEGs: differentially expressed genes; TUD: tobacco use disorder; DGE: differential gene expression.
See code for script summary.
See processed-data/SupplementaryTables to access and see the description of all supplementary tables generated in this study, including GitHub permalinks to the script lines in which they were created.
Original RNA-seq datasets that were used in this analysis are provided through the smokingMouse Bioconductor data package.
See raw-data access section for more details about this package contents and additional information about internal LIBD access.
Files are organized following the structure from LieberInstitute/template_project. Scripts include the R session information with details about version numbers of the packages we used.
- JHPCE locations:
/dcs05/lieber/marmaypag/smokingMouseGonzalez2023_LIBD001/dcl01/lieber/ajaffe/lab/smokingMouse_Indirects(old location)