Metatranscriptomic data(paired-end,Illumina)mapping?

[jojohan]

We have sequenced metagenomic data and metatranscriptomic data by Hiseq2000 paired-end of acid mine drainage.
we run velvet to assemble the metagenomic data,and we succeeded to assemble 3 draft genomes(or bining sequences).
I would like to map the metatranscriptomic data to 3 drafts. I have tried with bowtie and soapaliner/soap2.
soap -a RNA_1.fasta -b RNA_2.fasta -D draft_geneome_ORFS_db.index -v 3 -r 2 -s 40 -m 400 -x 600 -o PE.out -2 SE.out
soap -a RNA_1.fasta -D draft_genome_ORFs_db.index -v 3 -r 2 -s 40 -o SE.out
but the results are both very poor,the mapping results are less than 1% .
while I use the total contigs of metagenomic data as reference,the mapping result can be 83%.
Does anyone know the methods to deal with my problems? Thanks a millon!

[Jean]

So you are only mapping to the "3 draft genomes" you got from the assembly and ignoring the rest? How much of your genomic data is represented in these drafts? If you map back the metagenomic reads how much coverage do you get and how much of your data is leftover?

[jojohan]

So you are only mapping to the "3 draft genomes" you got from the assembly and ignoring the rest? How much of your genomic data is represented in these drafts? If you map back the metagenomic reads how much coverage do you get and how much of your data is leftover?

yes,because 3 drafts already make up the marjority of the metagenomic data. I haven't map back the metagenomic raw reads yet.I have another question about this. I will calculate the coverage of each drafts to both metagenomic raw reads and metatranscriptomic raw data. I have made some pretreatments of metatranscriptomic data including removal of low quality(average Q-value<20 per reads) and removal of reads with number of unknow bases (N) >10.so I am wondering if I should deal with metagenomic raw data in the excatly same way? Thanks!