Author ORCID Identifier

Year of Publication


Degree Name

Doctor of Philosophy (PhD)

Document Type

Doctoral Dissertation


Arts and Sciences



First Advisor

Dr. Stephen Randal Voss

Second Advisor

Dr. Jeramiah J. Smith


Among vertebrates, salamanders are champion regenerators, as they are able to regenerate a diverse set of tissues, and regenerate throughout their life. While it was originally thought that all salamanders regenerate similarly, recent studies comparing limb regeneration among deeply diverged salamanders (e.g. the axolotl and newts, ~150 million years) have identified deeply conserved commonalities, clear cellular and genomic deviations in the regeneration process. However, as salamanders belong to a speciose and morphologically diverse group, it is possible that cellular and transcriptional differences might also be present among closely related salamanders. Here, I performed a comprehensive transcriptomic analysis to rigorously identify transcriptional similarities and differences among three salamander species of varying divergences: A. mexicanum, A. andersoni, and A. maculatum. Most of these genes were associated with key regeneration processes noted in previous studies, suggesting that they are critical for the first 24 hours of limb regeneration. Unexpectedly however, genes responsible for tissue histolysis and musclespecific genes showed deviation in their regulation, suggesting that the transcriptional program for muscle histolysis varies among species that appear to complete regeneration at the same time. Further analysis of additional time points showed a similar result, suggesting that transcriptional networks deployed during limb regeneration may be evolutionary labile.

In addition, I developed and optimized a bioinformatics pipeline to investigate DNA methylation organization in the axolotl and better understand how this epigenetic mark is organized in a large, heavily repetitive, vertebrate genome. I used these methods, as well as methods to meta-analyze gene expression datasets of axolotl embryo tail regeneration, to identify significantly expressed regeneration genes that are differentially methylated during embryo tail regeneration process. The results from these analyses showed that many genes previously implicated as important during limb regeneration also exhibit significant changes in transcription and DNA methylation during embryo tail regeneration. This suggests that changes in gene expression and DNA methylation for these genes are linked and provides a starting point for future research in interrogating DNA methylation states and the expression of these genes during regeneration.

Digital Object Identifier (DOI)

Funding Information

The work presented here was supported by three fellowships given by the University of Kentucky: The Gertrude Flora Ribble Fellowship in 2015, The Biology Graduate Merit Fellowship in 2017, and the Morgan Graduate Fellowship in 2020. In addition, grants received from the National Institute of Health (R24OD010435, P40OD019794), from August 2015 - June 2021, and the Army Research Office (W911NF1010304, W911NF1410165), from August 2016 - June 2018, supported this work.

SupplementaryFile2.1.txt (1999 kB)
Significant alignments (> 98%) identified by BLASTn between V5 RNA-Seq contigs and V3 contigs that were used to design microarray probesets.

SupplementaryFile2.2.txt (272 kB)
The 2,360 A. mexicanum transcripts that were identified as commonly, differentially expressed using microarray and RNA-Seq, with RNA-Seq log2 fold changes. Additionally, A. andersoni and A. maculatum transcripts that significantly matched V5 contigs are reported, with RNA-Seq log 2-fold changes. Genes previously identified by Voss et al. 2015 are also noted, with log 2-fold changes. Transcripts that pass an FDR < 0.05 threshold in microarray and RNA-Seq, as well as in Voss et al. 2015 are noted.

SupplementaryFile2.3.xlsx (12 kB)
The list of 30 anonymous CRGs. Contig IDs identified from BLAST searches other A.mexicanum assemblies (Bryant et al. 2017, Caballero-Perez et al. 2018, Nowoshilow et al. 2018) are reported with V5 contig IDs. For anonymous transcripts, average coding potential calculated by CPC and sequence statistics (RNA and ORF size, peptide length) output from CPAT are reported.

SupplementaryFile2.4.xlsx (3433 kB)
The list of 405 non-redundant, conserved gene identifiers with RNA-Seq fold change values calculated for each species and as an average across species. Transcripts that passed an FDR < 0.05 in all comparisons are also noted.

SupplementaryFile2.5.tif (1619 kB)
Expression profiles from Voss et al. 2015 for 23 metabolism associated genes that were significantly downregulated at 24hpa.

SupplementaryFile4.1.xlsx (12 kB)
FastQC statistics for A. maculatum, A. andersoni, and A. mexicanum FastQ files. Columns include sample name, organism, number of raw reads prior to FastQ trimming, number of reads after FastQ trimming, and the percent of reads retained after trimming.

SupplementaryFile4.2.txt (21674 kB)
BLAST associations between the A. mexicanum (AxREFv2), A. andersoni (AndREFv2), and A. maculatum (MacREFv2) de novo assemblies. Microarray probe associations, Trinotate UniProt, Trinotate gene ID, AxPQv6 genome annotation, and final gene annotation is reported in columns D-H, respectively.

SupplementaryFile5.1.xlsx (2868 kB)
Differentially expressed genes (DEGs) identified comparing regenerating axolotl embryo tails vs. the initial control timepoint. Fold change values (log2) for each timepoint comparisons are noted in columns B-E, while false-discovery rate values are noted in columns F-I.