How to scale up library prep for large quantity of input DNA?

[spawn]

I am screening a library of 10-15k variants through cell-based assays. After the selection, I perform the NGS sample prep using NEB ultra II kits. When I try to validate any potential hits from the data, they all turn out to be false positives.

One of the things that I think might be going wrong is that I have a lower-fold coverage of DNA copies in my input DNA compared to the total library diversity. The recommended fold-coverage is 500-1000x i.e. for a library of 10k, I will need to process DNA from 10^3 x 500 = 500k cells, which comes to about 10-15 µg of DNA. The NEB kits have a limit of 1 µg per reaction. It becomes unaffordable when I have multiple samples and replicates. So I am looking for a way to scale up these reactions with alternative kits or methods. Any ideas?

Thanks,

[vivianshaw]

why not try Creative Biolabs' DNA Library Prep service?

[Ben3]

spawn could you just do a targeted approach to focus on the regions you're interested in? I'm not exactly sure which regions you're focused on, but I would think that building a comprehensive hybridization capture panel would work well. Many companies will do the design for you and some let you do the design yourself. I understand it might not be feasible depending on the scale of your project. Can you explain a little more about what you're sequencing?

[spawn]

Ben3 Thanks for responding. Sorry, I should've been more clearer in my original post. Basically I am interested in mapping junction points between the genomic DNA and transposes-mediated inserted foreign DNA. To do that, I fragment the genomic DNA using sonication, followed by end-repair, dA-taiing and ligation of adapters. I use the NEB Ultra II kit for this step. Then I amplify the junction regions utilizing primers that bind to ends of the inserted DNA and the ligated adapter. I am limited by the input quantity as I mentioned earlier--the kit only allows up to 1 µg of DNA per reaction and I probably want between 100-200 reactions, which becomes costly. I will look into the approach you mentioned.