Question: converting paired end reads to single end, is this a good idea?
1
gravatar for sciencer
14 months ago by
sciencer70
sciencer70 wrote:

Hi dear one,

I have paired end data and single end data (both are different single cell data), I want to find differential gene expression between them! Is this a good idea to convert paired end read to single end read? and How? Thanks in advance! Looking for your help, I am new to this field!

sequencing rna-seq next-gen • 1.3k views
ADD COMMENTlink modified 14 months ago by Asaf5.3k • written 14 months ago by sciencer70
2

Why do you want it at the first place? There are tools which will fairly work on paired data.

ADD REPLYlink written 14 months ago by bioExplorer3.7k

I want to find the differentially expressed genes between two data sets ( one is paired end and another is single end)

ADD REPLYlink written 14 months ago by sciencer70

As long as you remember to assign fragments to genes (with your paired-end data) rather than reads (default for single-end data) you should be okay to compare the two data sets.

ADD REPLYlink written 14 months ago by James Ashmore2.6k
1

If the only difference between the libraries is sequencing then you should be fine but I would just take read1 and ignore the other side, it might influence counting mapping to genes. However, I suspect that some other factors are different between these sets of libraries, e.g. fragmentation, selection, library generation kit and in this case your comparative analysis might find these biases instead of real biology.

ADD REPLYlink written 14 months ago by Asaf5.3k

Thank you Asaf........

ADD REPLYlink written 14 months ago by sciencer70

thank you but sorry, I didn't understand your answer *to assign fragments to genes * , can you please explain this, @James ashmore

ADD REPLYlink modified 14 months ago • written 14 months ago by sciencer70
2

A single-end read comes from sequencing one end of a fragment. A paired-end reads comes from sequencing the same fragment twice (each end of the fragment is sequenced). With single-end data, a single read represents a single fragment, but with paired-end data two reads (each pair) represents a single fragment. When you count how many reads align to each gene you will on average get twice the read count using paired-end data then you would using single-end data.

ADD REPLYlink modified 14 months ago • written 14 months ago by James Ashmore2.6k

Yes, thanks for your response

ADD REPLYlink written 14 months ago by sciencer70
1
gravatar for Devon Ryan
14 months ago by
Devon Ryan88k
Freiburg, Germany
Devon Ryan88k wrote:

No, as a general rule there's no need to convert your reads to single-end, aligners work just fine with paired-end data. The only exception is if you know a priori that the reads should overlap, in which can you can use flash or bbmap to overlap them to gain a small bit of speed (this tends to be more effort than its worth).

ADD COMMENTlink written 14 months ago by Devon Ryan88k

I want to find the differentially expressed genes between two data sets ( one is paired end and another is single end) , Is it possible?

ADD REPLYlink written 14 months ago by sciencer70

You have a batch effect to account for there, so even if you ignored read 2 from the PE dataset you still have that to account for. Is it possible? Yes, but it tends to be messy and the results are typically questionable.

ADD REPLYlink written 14 months ago by Devon Ryan88k
0
gravatar for Asaf
14 months ago by
Asaf5.3k
Israel
Asaf5.3k wrote:

As said before, there is no reason to do that. If you want to do it to compare read counts from different experiments, you might want to rethink this.

ADD COMMENTlink written 14 months ago by Asaf5.3k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 905 users visited in the last hour