problem for velvet parameters and results

[hequn]

Hi,everyone
I used velvet to assembly my data of transcriptome days ago,when it was finished,I get a file "contigs.fa" of 230M.I think it is too small.I use the parameters like "velveth output_directory 31 -shortPaired -fastq read1.fq read2.fq" and "velvetg output_directory -ins_length 200".Could you give me some advise?Thanks very much.

When I use velvet,should the pair-end read be same length? Since I have trimmed the reads for low quality,so the reads are different in the length size.

[bioBob]

Hi hequn.

could you give some statistics on your reads? Length and count? Also, how did you trim your reads? Did you make sure that the files are still in paired order? IE if you delete a read in file 1, did you delete it in file 2?

I do not think the members of the pairs have to be the same length with the one exception that they won't be considered if they are shorter than your kmer.

[hequn]

Hi bioBob,
Thanks for your answers.I haves nearly 30millions reads of 100bp,I trimmed the reads with the parameter "-a(quality score) 25 -l 25(default)" by using Btrim.It will remain the reads whose length is larger than 25bp.The Btrim can distinguish the single-end and pair-end reads automatically when it is finished.Today I tried k=21 and got a file "contigs.fa" of 253M.