Tweet
Hi,
I have RNA seq data which contain information about mRNA levels of two neuronal types (Eg: type A , type B).
Experiments,library preparation, sequencing and other downstream analysis to obtain FPKMs was done in a similar way.
As we lack other interesting neuron (type C) RNAseq data, my supervisor advised me to use Published data which obviously is of different protocol and method than ours. Published data on GEO has raw data and already preprocessed information containing the FPKM values of the particular neuronal sample I am interested in.
My question/ concerns:
Unique genes between samples:
1) From each sample, took the genes which satisfy the criteria FPKMS > threshold (along with statistical tests) and obtained sample specific expressed genes. Now I wanted to check how many unique genes exists between samples by excluding the common ones:
Question : Is it OK if I directly compare the genes list between samples??
2) Or Should I have to calculate FPKMs again from rawdata or do any kind of normalization in order to compare between datasets??
I have RNA seq data which contain information about mRNA levels of two neuronal types (Eg: type A , type B).
Experiments,library preparation, sequencing and other downstream analysis to obtain FPKMs was done in a similar way.
As we lack other interesting neuron (type C) RNAseq data, my supervisor advised me to use Published data which obviously is of different protocol and method than ours. Published data on GEO has raw data and already preprocessed information containing the FPKM values of the particular neuronal sample I am interested in.
My question/ concerns:
Unique genes between samples:
1) From each sample, took the genes which satisfy the criteria FPKMS > threshold (along with statistical tests) and obtained sample specific expressed genes. Now I wanted to check how many unique genes exists between samples by excluding the common ones:
Question : Is it OK if I directly compare the genes list between samples??
2) Or Should I have to calculate FPKMs again from rawdata or do any kind of normalization in order to compare between datasets??