How does degPatterns() decide what height to use for cutting tree

[sang-15]

Hi Lorena,

Thank you for providing this great package! I am just wondering how does degPatterns() decide what height to use for cutting the tree? And is there a way to set the height manually?

Also, in the normalized data generated from degPatterns(), there are a lot of columns named with cutoff, what are these columns and are these used during the decision making process?

colnames(patterns$normalized) [1] "genes" "merge" "value" "Animal.ID" "Age"
[6] "Sex" "Region" "Sequencing.run" "Analysis.ID" "Age_days"
[11] "colored" "cluster" "cutoff0.018" "cutoff0.071" "cutoff0.16"
[16] "cutoff0.284" "cutoff0.444" "cutoff0.64" "cutoff0.871" "cutoff1.138"
[21] "cutoff1.44" "cutoff2.151" "cutoff3.484" "cutoff4"

Thanks, Sang

[lpantano]

Hi, when running the function is going to use diana package to cut the tree. By default, several options are available, you can visualize these plots benchmarking and benchmarking_curve from the object that degPattern returns. You can choose a cutoff from there and use the function degPlotCluster to choose that cutoff to split genes with the option cluster_column indicating the value from the benchmarking plots that fits better your sample. You could do your own cutoff and then create a column in the table used by degPlotCluster if you want to plot the clusters in the same way.