the N50 is so low from soapdenovo

[heiya]

Hi， i used soapdenovo for my rice genome assembly，i tried many times，but the N50 is only 920(not an encouraged result it seems ) i also tried velvet, but it keep running for two week.
raw data: PE of reads length 90bp,20*coverage, insert size 500bp(only one library).
/soapdenovo63kmer all -s rice.config -o rice -k (31..59) -R -d -F -L 100

is there something wrong with my parameter set or is there any other tool for large genome assembly?

all reply is appreciated!

heiya

[flxlex]

With only 20 x coverage of 90 bp PE reads, this is about the best assembly you can expect. You would need at the very least to double the coverage, and preferably add Mate Pair reads as well (to get scaffolds)

[heiya]

With only 20 x coverage of 90 bp PE reads, this is about the best assembly you can expect. You would need at the very least to double the coverage, and preferably add Mate Pair reads as well (to get scaffolds)

thanks, flxlex!

[lovenlong]

Hi， i used soapdenovo for my rice genome assembly，i tried many times，but the N50 is only 920(not an encouraged result it seems ) i also tried velvet, but it keep running for two week.
raw data: PE of reads length 90bp,20*coverage, insert size 500bp(only one library).
/soapdenovo63kmer all -s rice.config -o rice -k (31..59) -R -d -F -L 100

is there something wrong with my parameter set or is there any other tool for large genome assembly?

all reply is appreciated!

heiya

Hi, heiya,
Have you find the way to improve your assembly's N50 ?
I'm dealing with the same illumina 101 PE data looks the same to yours.
Before assembly, quality filtering, trimming and errors correction would be essential.
I'd preprocessing and got assembly contig N50= 2.5Kb and scaffold N50=15Kb with Soapdenovo2 (Ave_ins 232, -K50).
But I'm wondering if the assembly can be done better.
Additionally, by mapping reads to contig sequence, the estimated average insert size was 232 +/-63 bp, looks different from the expected size (~ 360 bp)
Thankx