de novo assemly of Oases resulted in very high length transcript

[RyNkA]

As title, I was trying to do a multi kmer approach of Velvet/Oases on RNA seq data (single end read)

So I run Velvet and then oases, what I found out is that the transcripts.fa final file consist of a few contigs like this example

>Locus_1_Transcript_62758/62767_Confidence_0.018_Length_270364
GCGGCGTTCACCACTAGTCTTCCTCGCAAGCGCTCTGGCGTGGGCCTCCTTCTTTCAACT
TCATGGAAGGTGATGAGGTGGTCATAGTCAAAGGTATTGTCGATGTCTCTAACTTCATGG
AAGTCCATCCTGTCCAACCTGTTCCATCCTGGCGAGATCTTCAGCACAGCGTCGTCCTCC
TGAATGTGCTGTCTCTCATGTCGTTCGATCCTACCGCTGAGACCTAGACACAACCTTAGT
ATGTGATCAGCCTCTGACCTTTGGAGGAATTGAGGTGCTTATAGGCATGGTCCCTTGATA
AGAGACCGGTATCTCCAGCATATCCTTGGCATGATCCGCGTTCATGAAGAGACGGTATCC

So a transcript with a length of 270.000bp ? It also contains lots of N's inside. (through velvet final file *contigs.fa as well as the raw data do not have any N's)

*** I read somewhere on the seqans and found out that the oases algorithm try to do scaffolding even if there is no pair end read is available, that maybe the cause to the N's in the result transcripts.fa file

Any suggestions ?