Entering edit mode
6.5 years ago
Sreeraj Thamban
▴
290
Hello Biostars, I have a time course experiment RNASeq data,one strain of bacteria and three time points 0h,1h,and3h. Which is the best method to identify the expression pattern of genes over time using this data? I want to generate a image like this
You can try using http://bioconductor.org/packages/release/bioc/html/EBSeq.html EBseq package - will generate all the possible patterns between various conditions,and gives you the results in terms of genes following that pattern.
Hello Ron, I checked the EBseq package, but I am bit confused about interpretation the 'pattern' which is given by EBseq. It gives a list of patterns but what exactly those pattern means?
The pattern means if you have for e.g. 3 classes. A gene can go up in first,up in second class and down in the 3rd class.This is one pattern. Also it could go up in first,up in second and up in third.this is another pattern. Similarly you can have all kinds of combinations and see which pattern is of your interest. Up-down- Up or Down-Down-Up ?
What is the format of your data? If it's raw count data you could use DEseq2 with time-series design. WIthout DE testing, uou can generate profiles like the one you linked by clustering the normalized count data.