line graph

[zluo001pfr]

Hi Dr Li,

I am trying to use your code for my RNA-seq data and I am up to the line graph part. I have a graph here but it looks a bit odd and seems the lines jumped between my tissues (bud2, 3, 6, 7). I am hoping you might have some suggestions about what I have done wrong.

[cxli233]

When this happens, usually it is because the replicates are not averaged, or not averaged properly. Can you make sure every gene only has 4 observations (bud2, 3, 6, 7)? If you have more than 1 observations per time point (?), it leads to "jumpy" time points. Please review the code chunk where you average to reps. Try removing library or sample name from the group_by() step.

[zluo001pfr]

thanks for the super quick reply. I will check my code. thank you very much. while you are here, sorry. more questions.. i have plot the edge table but it looked different from yours. is this ok?

[cxli233]

It seems all your genes are either positively correlated with each other, or anit-correlated with each other. How many genes did you put into the correlation? This is an indicator of not having enough genes correlation. I think this is fine... If you are getting results that make biological sense, I wouldn't be too worried about it.

[zluo001pfr]

thanks again. I used the high var (top 40%) and high F (F>2.5) genes which are about 8600 genes from a dataset that has >30k genes.

[cxli233]

That seems fine. I never seen patterns like this, but I think it should be okay.

[zluo001pfr]

cool, thanks

[zluo001pfr]

When this happens, usually it is because the replicates are not averaged, or not averaged properly. Can you make sure every gene only has 4 observations (bud2, 3, 6, 7)? If you have more than 1 observations per time point (?), it leads to "jumpy" time points. Please review the code chunk where you average to reps. Try removing library or sample name from the group_by() step.

Hi Dr Li, me again. I should mention that I have 3 factors in this exp, position (x4, bud 2,3,6,7), timepoint (x2, 3h and 24h), and treatment (x2, high and low). so maybe that's why I have this issue. I group_by(gene, position, time_point, treatment). but i graph using just two factors.

[cxli233]

This is exactly why you have this issue. Try making the same graph but separately for high vs low treatment. You can also try facet_grid(time_point+treatment~module). You will need to make the graph longer/taller.

On Thu, Oct 6, 2022 at 17:18 Luke Luo ***@***.***> wrote: When this happens, usually it is because the replicates are not averaged, or not averaged properly. Can you make sure every gene only has 4 observations (bud2, 3, 6, 7)? If you have more than 1 observations per time point (?), it leads to "jumpy" time points. Please review the code chunk where you average to reps. Try removing library or sample name from the group_by() step. Hi Dr Li, me again. I should mention that I have 3 factors in this exp, position (x4, bud 2,3,6,7), timepoint (x2, 3h and 24h), and treatment (x2, high and low). so maybe that's why I have this issue. I group_by(gene, position, time_point, treatment). but i graph using just two factors. — Reply to this email directly, view it on GitHub <#1 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALLXJPDGYHKJG6PTMQLSHH3WB462RANCNFSM6AAAAAAQ6624AI> . You are receiving this because you modified the open/close state.Message ID: ***@***.***>

-- Li

[zluo001pfr]

thank you. fixed it. 👍