abyss-pe output after filtering reads

Question

[max_19]

Hi all,

I have a genome assembly pipeline which involves identifying reads which map to contaminants such as bacteria. I eventually remove these reads and re-assemble again. My question is regarding the output of abyss pre and post filtering of these reads. The genome I am currently processing is mostly contaminated with bacteria, below is the scaffold stats of abyss

                 N                N:500          E-size         Sum
Original         81               54             256349         4619298
Post-filter      7621             2006           1226           1805249

Code I used to run abyss

> abyss-pe k=96 name=novo in='R1.fastq R2.fastq'

Does anyone know why N would increase so drastically? Also, I found this in the documentation for E-size: E-size=sum of the square of the sequence sizes divided by the assembly size, but i'm not sure what the biological interpretation of this value is?

any help would be so appreciated! Thank you.