How to determine the insert length?

[anyone1985]

I have some Solexa pair-end data. But my colleague forgot to tell me the insert length How can I determine the insert length of the data? First, I have no reference genome.

[westerman]

The easiest way would be to go back to your colleague and get the insert length. But if you really insist on doing this the hard way then I would suggest doing a de-novo assembly using the reads as if they were fragments (i.e., not as paired ends). This should give you some decent size contigs. Then map the paired ends map onto the contigs. From this you should be able to figure out how far apart are the paired ends that do map. After you obtain the numbers that define the range you can then do an new assembly but this time as a 'paired end' instead of a 'fragment' assembly.

[anyone1985]

thank you, i think i know what i should do

[system7]

I have some Solexa pair-end data. But my colleague forgot to tell me the insert length How can I determine the insert length of the data? First, I have no reference genome.

The summary.htm file from Pipeline should have that info in it at the bottom of the file

[Torst]

The easiest way would be to go back to your colleague and get the insert length.

The problem with this is that the DNA fragment selection step is inexact. You may be aiming for 250 bp, but the average is 220 say, with a standard deviation of 30.

But if you really insist on doing this the hard way then I would suggest doing a de-novo assembly using the reads as if they were fragments (i.e., not as paired ends). This should give you some decent size contigs. Then map the paired ends map onto the contigs. From this you should be able to figure out how far apart are the paired ends that do map. After you obtain the numbers that define the range you can then do an new assembly but this time as a 'paired end' instead of a 'fragment' assembly.

This is good advice. If you have a close reference sequence, you can use that instead of de novo contigs. I usually use MAQ to align a SUBSET of the reads in paired-end mode, and MAQ itself will print out the mean and s.d. of the insert size.

And as another poster said, if this is Illumina GA Pipeline, the Summary HTML files contain an estimate of the insert size which it obtains by using ELAND to map the reads to the reference genome specified in the gerald.cfg file.

[westerman]

The problem with this is that the DNA fragment selection step is inexact. You may be aiming for 250 bp, but the average is 220 say, with a standard deviation of 30.

Well yes this could be a problem if your colleague only gives a single number then you have problems. I always ask for a minimum and maximum insert length knowing that those numbers are also uncertain. Also sometimes you can have a mixture of libraries with different insert sizes; e.g. average of 500 bp; 3K, 20K. Then one needs to know not only the range but also which reads corresponds to which library.

And as another poster said, if this is Illumina GA Pipeline, the Summary HTML files contain an estimate of the insert size which it obtains by using ELAND to map the reads to the reference genome specified in the gerald.cfg file.

Ah, but the original poster said he did not have a reference genome.

It was an interesting theoretical question -- how does one figure out insert sizes when only given paired ends. A question that I am glad that I do not have to do in practice!

[ashrafi_h]

How do you use maq to determine the insert size?

The problem with this is that the DNA fragment selection step is inexact. You may be aiming for 250 bp, but the average is 220 say, with a standard deviation of 30.



This is good advice. If you have a close reference sequence, you can use that instead of de novo contigs. I usually use MAQ to align a SUBSET of the reads in paired-end mode, and MAQ itself will print out the mean and s.d. of the insert size.

And as another poster said, if this is Illumina GA Pipeline, the Summary HTML files contain an estimate of the insert size which it obtains by using ELAND to map the reads to the reference genome specified in the gerald.cfg file.

Hi, If you do not have ref sequence, how do you use maq to determine the insert size. Could you please have a sample command line.

Thanks

[ashrafi_h]

The summary.htm file from Pipeline should have that info in it at the bottom of the file

Where is this summary.htm that people are talking about?