MARTEENI compiled and processed sample data

Published: 27 February 2021| Version 1 | DOI: 10.17632/ktg5fp9gzn.1
Contributor:
Mary Kasper

Description

Compiled data sets that include averaged sample data. Raw data can be found at http://dx.doi.org/10.17632/4s45ntxhw3.1

Files

Steps to reproduce

All statistical analyses were performed using JMP Pro 14 and R-Studio. Statistical differences between experimental groups were determined with 2-way ANOVA for mechanical and in vivo intensity measurements (i.e., S100, collagen, laminin), 3-way ANOVA for in vivo axon density and fibrotic capsule measurements, and 1-way ANOVA for axon diameter. In vitro migration data were initially plotted as histograms to examine their distribution and both Poisson and negative binomial distributions were identified as suitable representations for the counts of cells. Stepwise selection was used to compare potentially predictive models, while their fits were compared using likelihood-ratio tests and the Akaike information criterion in R-Studio. A 2-way generalized linear mixed model following the negative binomial distribution was identified to best predict cell counts, especially when concentration of the hydrogel, templating process, and their interactions were considered as "fixed effects", whereas discrete depth locations were assumed as "random effects". To further examine the levels of discrete depth bins every 50 µm, data from templated gels only were fit to a 2-way generalized linear mixed model. Hydrogel concentration and depth were inputted as "fixed effects", while sample levels were inputted as "random effects". Analyses were followed by Tukey’s honestly significant difference posthoc test for multiple comparisons, with an overall confidence interval of 95%. In vitro migration data are graphed as minimum observation, lower 25% quartile (Q1), median, upper 75% quartile (Q3), and maximum observation. All other data are represented as the mean ± standard deviation (SD).