Hi all,
I have recently tried using Velvet Columbus extension in order to preform reference guided assembly of a whole genome sequence.
I have tried using two references to which I mapped my read using BWA -
The better one had 99.57% of reads mapped (and a cover of 99.9% of the reference)
The worse one had only 84.54% of the reads mapped.
I tried assembling with velvet Columbus with both of the references (the rest of the parameters were the same).
Surprisingly, the worse reference had better result (225 contigs, L50 = 29) than the better reference (370 contigs, L50 = 39)
My question is, are those results plausible? I used velvet running on the BAM file produced by BWA with the same parameters exactly. Is there anything wrong that I could have done on the way?
Thank you a lot!!
I have recently tried using Velvet Columbus extension in order to preform reference guided assembly of a whole genome sequence.
I have tried using two references to which I mapped my read using BWA -
The better one had 99.57% of reads mapped (and a cover of 99.9% of the reference)
The worse one had only 84.54% of the reads mapped.
I tried assembling with velvet Columbus with both of the references (the rest of the parameters were the same).
Surprisingly, the worse reference had better result (225 contigs, L50 = 29) than the better reference (370 contigs, L50 = 39)
My question is, are those results plausible? I used velvet running on the BAM file produced by BWA with the same parameters exactly. Is there anything wrong that I could have done on the way?
Thank you a lot!!
Comment