Time course RNA seq with DESEq2
1
0
Entering edit mode
23 months ago
synat.keam ▴ 100

Hi all,

Hope you are well. I am running time course experiment to examine differentially expressed genes from tumour tissues between two conditions, radiotherapy and sham-radiotherpapy groups at 4 time points post radiotherapy. Each time point had size matched untreated controls. I used the following code to run model with interaction.

Option 1

dds_all <- DESeqDataSetFromMatrix(countData = Raw.Count, 
                               colData = Phenotype,
                               design = ~ Treatment + Time + Treatment:Time)

set reference group

dds_all$Treatment <- relevel(dds_all$Treatment, ref ="Sham-radiotherapy")

DESeq

dds_all <- DESeq(dds_all)

Examine result

resultsNames(dds_all)

all_time_interaction <- results(dds_all, alpha =0.05)

summary(all_time_interaction)

with this model, I got around 100 differentially expressed genes (not many)

Additive model

dds_additive <- DESeqDataSetFromMatrix(countData = Raw.Count, 
                                   colData = Phenotype,
                                   design = ~ Treatment + Time)

relevel

dds_additive$Treatment <- relevel(dds_additive$Treatment, ref ="Sham-radiotherapy")

DESeq

dds_additive <- DESeq(dds_additive)

Examine result

resultsNames(dds_additive) 

all_time_additive <- results(dds_additive,alpha = 0.05)

summary(all_time_additive)

Additive model (no interaction) gave me around 3000 differentially expresse genes.

Hard to make a decision whether to retain the interaction or drop it because I did not see any p-value for interaction term? Does anyone knows where I can I look whether interaction is signficant? so I know whether I should keep it in the model or not? what is your suggestion to move forward!

Option 2

I found tutorial using the following command to analyse Time course RNA seq using DEseqDataset function. I then tried it with my dataset using following command.

Time_course_RT <- DESeqDataSet(Raw.Count, ~ Treatment + Time + Treament:Time)

Time_course_RT <- DESeq(Time_course_RT, test="LRT", reduced = ~ Treatment + Time)

My question is whether my option 1 using DESeqDataSetFromMatrix is correct to analyse Time course RNA seq compared with DEseqDataSet?

I got error once I used DEseqDataSet with the following message

*Error in DESeqDataSet(Raw.Count, ~Treatment + Time + Treament:Time) : 
  'se' must be a RangedSummarizedExperiment object*. 

How do I fixed it? how could I convert my data into RangedSummarizedExperiment object ?

I found some people experienced this issue too, but did not get any definitive answer. Thank in advance for your help.

Regards,

Synat,

DESeq2 • 1.1k views
ADD COMMENT
0
Entering edit mode
23 months ago

Running results without specifying a contrast or a ResultsName is dangerous, are you absolutely sure you know what results you are getting?

ADD COMMENT
0
Entering edit mode

HI Swbarnes,

This is time course examining genes that are differentially expressed overtime, not at specific time point. So, in this case, I have only two groups, radiotherapy and sham-radiotherapy collecting at different time points and Time variable is now included in the model and also Time:Treatment interaction. I have specified reference group which is sham-radiotherapy so the comparision is between radiotherapy vs. sham-radiotherapy adjusting for time or time:treatment interaction. I used to specified contract previously, but setting reference level is enough for two groups. I have checked already and the model did compare between radiotherapy_vs_sham-radiotherapy. Hope i am right!

Thanks, Synat

ADD REPLY

Login before adding your answer.

Traffic: 3224 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6