Nov. 3-5, Salmon DMR analysis

Find DMRs

• I updated my DMR analysis by including reads with MAPQ score of >= 20
• Using methylpy pipeline, I
  • created allc files
    • mox script here: 20191030_DMRfindMAPQ20_Salmo.sh
    • allc files here: https://gannet.fish.washington.edu/metacarcinus/Salmo_Calig/analyses/20191030/
  • created filtered bam files for QC
    • mox script here: 20191030_DMRfindMAPQ20_SalmoPt2.sh
    • filtered bam files here: https://gannet.fish.washington.edu/metacarcinus/Salmo_Calig/analyses/20191030/
  • ran DMRfind to identify within sample DMRs that show methylation at a significantly different level than would be expected by genetic variation alone
    • mox script here: 20191030_DMRfindMAPQ20_SalmoPt2.sh
    • DMR results here: https://gannet.fish.washington.edu/metacarcinus/Salmo_Calig/analyses/20191030/

Filter DMRs

• filtered DMRs for those showing >=5x coverage in 3/4 samples per group. I didn’t use this cutoff for the control samples since these each only has two individuals.
  • R script here: DMR_cov_in_0.75_SamplesPerCategory.Rmd
  • R project here: DMR_cov_in_0.75_SamplesPerCategory.Rproj
  • filtered DMRs here: DMR250bp_MCmax25_cov5x_rms_results_filtered.tsv

Group stats on DMRs

• R script: MCmax25_asinT_groupStats.Rmd
• R project: MCmax25_asinT_groupStats.Rproj

Data preprocessing:

• To determine if there is a significant experimental effect on DMRs I first arcsin(sqrt) transformed the percent methylation data
• BEFORE transformation percent methylation distribution:
• • AFTER transformation percent methylation distribution:

Run ANOVA

1. I ran a 1 way ANOVA to test if sea lice infestation had an effect
   • 1 way ANOVA results summary table here: DMR_MCmax25_1wayAOV_infest_modelsumm.csv
   • 3 DMRs showed a significant infestation effect at 1way ANOVA p-value of < 0.01
   • 13 DMRs showed a significant infestation effect at 1way ANOVA p-value of < 0.05
   • 24 DMRs showed a significant infestation effect at 1way ANOVA p-value of < 0.1
     Heatmap of 24 DMRs that showed a significant infestation effect at 1way ANOVA p-value of < 0.1. Heatmap key:Column color bar in heat maps below: light pink = 16C_26psu, dark pink = 16C_32psu, light green = 8C_26psu, dark green = 8C_32psu, magenta = CTRL_16C_26psu, cyan = CTRL_8C_26psu. heatmap color: Red = more methylation, blue = no methylation, black = no data.
2. For the 24 DMRs that showed a significant infestation effect at 1way ANOVA p-value of < 0.1, I ran a two way ANOVA on their percent methylation to see if temperature, salinity, or their interaction showed a significant effect.
   • 2 way ANOVA results summary table here: DMR_MCmax25_2wayAOV_TxS_modelsumm.csv
   • Tukey HSD results summary table here: DMR_MCmax25_DMR_TukHSD_TxS_modelsumm.csv
   • 3 DMRs showed a significant Salinity effect (at adj.p.value < 0.1)
   • 3 DMRs showed a significant temperature effect (at adj.p.value < 0.1)
   • 6 DMRs showed a significant temperature effect (at adj.p.value < 0.1)

Next steps

• validate 24 DMRs that showed a significant infestation effect at 1way ANOVA p-value of < 0.1 in IGV
  • generate ANOVA filtered bed file
  • compare to filtered bam files linked above
• find out where in the genome these DMRs are