TPM usage instead of FPKM in ballgown sattest

[Jiglov]

Hi, there I've got to the point of DE expression analysis with ballgown, after creating ballgown object.

However after reviewing ballgown documentation I did not find any reference to use TPM instead of FPKM https://github.com/alyssafrazee/ballgown

And when trying to use TPM I get an error, RESULTS = stattest(bg_gene_filt, feature = "transcript",covariate = "population", meas="TPM") Error in match.arg(meas) : 'arg' should be one of “cov”, “FPKM”, “rcount”, “ucount”, “mrcount”, “mcov”

This did work for FPKM:

RESULTS = stattest(bg_gene_filt, feature = "transcript",covariate = "population", meas="FPKM")

1. I don't want to use covariate, but do want to use adjustvars = "population" When I remove covariate I get error message: RESULTS = stattest(bg_gene_filt, feature = "transcript",adjustvars = c("population"), getFC = TRUE, meas = "FPKM") Error in stattest(bg_gene_filt, feature = "transcript", adjustvars = c("population"), : invalid covariate name

2. I also want to get FC getFC = TRUE, gives In stattest(bg_gene_filt, feature = "transcript", covariate = "population", : fold changes only available for 2-group comparisons

Basically I want to compare 3 groups of 4 replicates each and get FC, p and q values, and if possible FDR. But all this for TPM and not FPKM....

Thank you very much for the previous help, appreciate any suggestions here...

[JMF47]

Ballgown currently does not support using TPM, but only to the list of measures: 'arg' should be one of “cov”, “FPKM”, “rcount”, “ucount”, “mrcount”, “mcov”. The package is not currently actively developed, but I can see if I have some time to incorporate TPM at some point.

[Jiglov]

Thanks, that would be great, if you could find some time to incorporate TPM in...

[xingjianfeng100]

Thanks, that would be great, if you could find some time to incorporate TPM in...

Should there be a difference between TPM and FPKM? why insisted on using TPM?