Some suggestions for optimizing TCseq

[nzx9581]

Hello developer,

Thanks for providing such a great package for time-course analysis of ATAC/ChIP-seq, and I did use this package in my project deeply. I have a suggestion for optimizing this package and I hope that will help you to keep providing people with good functions.

I think it's a "little bug" in your code but could cause severe consequences. Normally function DBanalysis works well if users input the timepoint info as characters, such as "4 years" or "3 days", it's fine, and the software works very well. But if I provide timepoint info just like "11" or "72" which lack of time unit, the error information "Error in design[, unique(group)] : subscript out of bounds" happened.

I checked your code, you used design <- design[, unique(group)] to extract certain columns for the next step of analysis. So that's the problem, if users provide the timepoint information as "4 years", it won't have any problem because they are characters and R will extract columns based on the column name. But if users provide timepoint as "71", R will directly pick up columns based on the column number instead of the column name or column index because R deems "71" as the column number.

If someone provides their timepoint info without any characters just numbers, it will leads to error information and if the number of timepoints is accidentally smaller than the total number of columns in design data frame, it will also cause severe mistakes that will mislead people's work. So I hope you can fix it in the next version. Thank you!

[nzx9581]

For people who encounter the error Error in design[, unique(group)] : subscript out of bounds when using TCseq, please check your timepoint column in your design dataframe and avoid providing time without any unit or letters.