I am new to DESeq2 . I have 394 unique patient RNA sequencing samples of ovarian primary tumor. After reading information about the package, I am still confused by the word , "replicates". In the context of my data, each RNA sequencing sample represents a unique patient's primary tumor, and DESeq2 would treat these as individual replicates of the primary tumor condition, correct? Therefore, I can conduct analysis on this data using this tool?
I want to use this perform normalization and differential expression of the molecular subtypes within my dataset.
I figured DESeq2 would be appropriate because I plan to normalize and compare another dataset to this given dataset later on, and wanted the same processing steps.
In the context of RNA-Seq there are two types of replicates: technical replicates and biological replicates. When you sample the RNA from different individuals these are called biological replicates; but, when you sample (more than one) the RNA (from your sequencing library) from the same individual these are called technical replicates.
So, don't worry about this if each sample comes from each independet person. But, take into account that when you have technical replicates you need to collapse these replicates. DESeq2 have a collapse replicates fucntion.
Ahh okay, so in this case I have biological replicates of the primary tumor condition. I have no technical replicates being used in this dataset.
I will take that into account for my other dataset though, it has technical replicates.
So we can perform Differential expression on the dataset using a clinical sheet (in this case I wanted to compare the expression analysis of the different molecular subtypes within my dataset)? I don't have normal vs tumor or something like that this for one.
Yes, you can compare two conditions (or subtypes). All you need is an experiment design: ~subtypes. Take into account that you need at least 3 replicates for each condition or subtype to have a better estimate for the statistics.
Ahh okay, so in this case I have biological replicates of the primary tumor condition. I have no technical replicates being used in this dataset.
I will take that into account for my other dataset though, it has technical replicates.
So we can perform Differential expression on the dataset using a clinical sheet (in this case I wanted to compare the expression analysis of the different molecular subtypes within my dataset)? I don't have normal vs tumor or something like that this for one.
Yes, you can compare two conditions (or subtypes). All you need is an experiment design: ~subtypes. Take into account that you need at least 3 replicates for each condition or subtype to have a better estimate for the statistics.
Perfect Piero. Thanks again, you are the best! Looking forward to conducting the analysis!