Question: ballgown time course design and command confirmation
0
19 months ago by

Hi, all professionals!

I have some confuse here in performing timecourse analysis of my samples. here is my expdesign, did I run a correct command for the time course analysis. Thank you all!

the header of the design matrix is ::

IDs               tissue       reps     time     timen
xxxx_L0-1    shoot         1            0         1
xxxx_L0-2    shoot         2            0         1
xxxx_L0-3    shoot         3            0         1
xxxx_L05-1    shoot         1            0.5       2
xxxx_L05-2    shoot         2           0.5       2
xxxx_L05-3    shoot         3            0.5       2


...

I have 12 time points. I want to timecourse analysis plus comparison in two tissues with ballgown, here is my R command, does it meet all my requirements?

tissue_adj_timecourse_transcript = stattest(bg, feature="transcript", meas='FPKM', covariate='time', timecourse=TRUE, adjustvars='tissue')

I want to compare the differences among the different time poits, and also the differences between two tissues (shoot and root) along the timecourse. Is this right? which covariate is better, 'time' or 'timen'? if I don't add the adjustvars, does the command mix all samples from the same time point (combining all reads or assemblies of shoot and root with 3 replicates from, say, time point 0.5h) and comparing to find the differences?

thanks again!

I tried to use 'timen' as the covariate, the transcripts and genes are a bit different with the two different covariate, 'time' and 'timen'.

btw: what type of "reps" is it? character or numeric? I loaded it as the character, I don't know if it is ok when doing analysing.

thanks!

written 19 months ago by yadongxue0

could any one illuminated me on this? thanks.