Newbler Metrics

[Narmneung]

Hi all,

I am currently trying to assemble a ~400 Mb plant genome using Newbler version 2.8. I have Illumina Hi-Seq paired-end reads (135 M reads), 90 bp with 500 bp insert. Below is the statistics:

//scaffoldMetrics
avgScaffoldSize = 50577;
N50ScaffoldSize = 162943, 320;
largestScaffoldSize = 3497334;

avgScaffoldContigSize = 4054;
N50ScaffoldContigSize = 7987, 7459;
largestScaffoldContigSize = 212004;

scaffoldEndMetrics
NoEdges = 9582, 91.2%;
OneEdge = 0, 0.0%;
TwoEdges = 0, 0.0%;
ManyEdges = 0, 0.0%;
LargeRep = 920, 8.8%;

scaffoldGapMetrics
BothNoEdges = 45419, 88.8%;
OneNoEdges = 5347, 10.5%;
BothOneEdge = 0, 0.0%;
MultiEdges = 0, 0.0%;
LargeRep = 356, 0.7%;

largeContigMetrics
numberOfContigs = 68739;
numberOfBases = 238994887;

avgContigSize = 3476;
N50ContigSize = 7521;
largestContigSize = 212004;

Q40PlusBases = 237415811, 99.34%;
Q39MinusBases = 1579076, 0.66%;

largeContigEndMetrics
NoEdges = 129091, 93.9%;
OneEdge = 0, 0.0%;
TwoEdges = 0, 0.0%;
ManyEdges = 0, 0.0%;
LargeRep = 8387, 6.1%;
//

Google tells me that "NoEdges" means the assembly is bad.
Any advices?

Thank you!

[flxlex]

Yep, this is not good - even though the meaning of these egde metrics is poorly documented. I would do a bunch of QCs on the reads (not just fastqc, also preqc from SGA) and assembly (concoct for example) to find out what is going on. The scaffold metrics are not too bad for a PE dataset only, by the way...