Hello experts,
I posted similar question last year but I haven't figured out the problem. I went through old posts too to see if I could find the answer but couldn't. Any suggestion or idea on my problem would be greatly appreciated.
So I have Illumina HiSeq 100 bp PE reads (library insert size is 200 bp). After QC of reads, I assembled reads using velvet. The average contig size is about 250 bp. Anyway, I tried to map the reads on the assembled contigs using bowtie and only less than 10% of reads were mapped on the assemblies. I just observed that "contig.fa" file generated from the velvet assembly contains very short length of contigs which are between 40 - 60 bp. I realized that I used read length cutoff of 10 bp during QC meaning I kept all reads longer than 10 bp. Maybe these short reads were used for the assembly and generated weird contigs resulting in low mapping percentage? This maybe also a reason for short average contig length?
I posted similar question last year but I haven't figured out the problem. I went through old posts too to see if I could find the answer but couldn't. Any suggestion or idea on my problem would be greatly appreciated.
So I have Illumina HiSeq 100 bp PE reads (library insert size is 200 bp). After QC of reads, I assembled reads using velvet. The average contig size is about 250 bp. Anyway, I tried to map the reads on the assembled contigs using bowtie and only less than 10% of reads were mapped on the assemblies. I just observed that "contig.fa" file generated from the velvet assembly contains very short length of contigs which are between 40 - 60 bp. I realized that I used read length cutoff of 10 bp during QC meaning I kept all reads longer than 10 bp. Maybe these short reads were used for the assembly and generated weird contigs resulting in low mapping percentage? This maybe also a reason for short average contig length?
Comment