Regarding paired -end assembly insert-size calculation on newbler2.3

[ganga.jeena]

pairedReadStatus
{
numberWithBothMapped = 1157826;
numberWithOneUnmapped = 229595;
numberMultiplyMapped = 715851;
numberWithBothUnmapped = 11362;

3KB insert library
library
{
libraryName = "GP5JY2V01.sff";
pairDistanceAvg = 2614.3;
pairDistanceDev = 653.6;
}

library
{
libraryName = "GP5JY2V02.sff";
pairDistanceAvg = 2608.8;
pairDistanceDev = 652.2;
}



8KB insert library
library
{
libraryName = "GSMQLLJ01.sff";
[B] pairDistanceAvg = 2719.4;[/B]
pairDistanceDev = 679.8;
}

library
{
libraryName = "GSMQLLJ02.sff";
pairDistanceAvg = 2702.5;
pairDistanceDev = 675.6;
}


Can someone explain the error in calulation of the 8kb library insert size in newbler v2.3 gsAssembler .
If it is due to use of same linker or on what basis is the paired-end distance calculated.

If we give 20kb library how will the system differ it from 3kb library data.

Is there any option in command line , whereby we can specify the insert size......that I am missing??


Regards,
Ganga Jeena

[Torst]

8KB insert library
pairDistanceAvg = 2719.4;
Can someone explain the error in calulation of the 8kb library insert size in newbler v2.3 gsAssembler .

I suspect, like many others, that your 8 kbp library was not prepared very well, and ended up as a 3 kbp one. You should ask the sequence provider about it.

[jurgen_]

I have a similar issue. I ran Newbler on a combination of a 454 single-end and an Illumina paired-end library with an insert size of 200 bp. The pairDistanceAvg in the metrics file is wrongly estimated to be 612 bp. When I manually calculate the average pair distance from all 'SameContig' rows in the 454PairStatus.txt I get an average of 212 bp and a standard deviation of 21 bp, which is more likely to be correct.


pairedReadStatus
{
numberWithBothMapped = 3653513;
numberWithOneUnmapped = 149804;
numberMultiplyMapped = 49574;
numberWithBothUnmapped = 471;

library
{
libraryName = "Illumina";
pairDistanceAvg = 612.0;
pairDistanceDev = 243.8;
}
}


The funny thing is that when I run Newbler on the 454 dataset plus a subset of 40% of the above mentioned Illumina data, the paired-end distance is estimated correctly to be 214 bp.