How Many PCR Reactions Do I Need For NGS Analysis Of Human Gecko v2 Screen?

[Yoshi]

Hi CRISPR experts,

First, I would like to thank the Zhang lab team and all of you for giving us an opportunity to discuss CRISPR.

I am going to do genome-wide knock-out screening with Human Gecko-v2.
I have a question about the required number of PCR reactions for amplification of gDNA before and after selection using 1step PCR method in"Nature protocol 2017".

My calculation for PCR reactions is here,

Coverage: 500 cells per sgRNA
Number of sgRNA in Library A: 65383
Required number of cells to achieve 500x : 33,000,000 cells

We will get approximately 220 ug of gDNA from 33 million cells(assuming 6.6ug of gDNA for 1 million cells in"sience 2014")
Each 50-ul PCR reaction can hold up to 2.5ug of gDNA (at the step 58 in "Nature protocol 2017").
Therefore, 88 PCR reactions (220ug divided by 2.5ug) will be required for each conditions and bioreps.

For example, if we have 2 conditions (treated cells and un-treated cells) and 2 bioreps at the end of screen (2x2) and a control condition before screen selection (1), 440 PCR reactions (=88x2x2+88) are required.
Human Gecko-v2 consists of 2 libraries (libraryA and libraryB). Maybe, at least 880 PCR reactions are required.
I could be wrong. For me, 880 PCR reactions are too many.
I appreciate if someone could give us some advice about how to handle PCR reactions that can amplify all the gDNA harvested from the screen.

Thanks,
Yoshi

[Julia Joung]

Hi Yoshi,

Yes you are correct - a genome-scale screening readout requires a large number of PCRs. Before setting up all of the gDNA PCRs, I would recommend running a test reaction to see how much gDNA the reaction can hold for efficient amplification. Different cell types can generate different qualities of gDNA that may affect the PCRs.

The one nice thing about screening readout is that you can make 1 mastermix for each screening condition, and aliquot the mastermix into different wells.

Best,

Julia

--
You received this message because you are subscribed to the Google Groups "Genome Engineering using CRISPR/Cas Systems" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crispr+unsubscribe@googlegroups.com.
To post to this group, send email to cri...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Chris McDermott-Roe]

That's a scary number of reactions. Thanks for running the numbers!

[Yoshihiro Gocho]

Hi Julia,

Thank you so much for your answer. When I see a lot of flasks during my CRISPR screening and calculate the number of PCR, I sometimes think I could do anything wrong.

 I have never put such a big amount of DNA like more than 1 microgram per reactions. I will run a test reaction to set up PCR.

If I would like to reduce the number of PCR, can I substitute 300X coverage or 100X coverage for 500X coverage? or can I amplify a part of DNA (not all the DNA harvested) ?

Positive selection with 300X coverage (or 100X coverage) may be OK?? Negative selection with 300X coverage (or 100 coverage) may not be OK???

Best,

Yoshi

--
You received this message because you are subscribed to a topic in the Google Groups "Genome Engineering using CRISPR/Cas Systems" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/crispr/Qk4V_oy82iM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to crispr+unsubscribe@googlegroups.com.

[Julia Joung]

Hi Yoshi,

It's true that positive selection requires a lower coverage than negative selection, but I'm usually a bit hesitant to reduce the coverage. You can try with 300X coverage for positive selection to see if you can get enough signal, but I would not necessarily recommend it. Given how costly NGS is, it is probably not worth the risk.

Best,

Julia

--

You received this message because you are subscribed to the Google Groups "Genome Engineering using CRISPR/Cas Systems" group.

To unsubscribe from this group and stop receiving emails from it, send an email to crispr+unsubscribe@googlegroups.com.

[Yoshihiro Gocho]

Hi Julia,

Thank you so much for your reply. I totally agree with you.

Best,

Yoshi

[athe...@gmail.com]

Hi Everyone,

Thank you so much for sharing your opinions here. I have a related question. Based on Yoshi's calculation, there are 88 PCR reactions required for each condition and biorep.  How to choose the primer sets for all these 88 reactions?  From Julia's paper, we need use 10 NGS-lib-Fwd primers and 1NGS-lib KO-Rev primer. Does this mean there are 10 reactions for each sample? I am confusing about this. Any help would be greatly appreciated.

Best,

Mei

[Yoshihiro Gocho]

Hi Mei,

Thank you for joining the discussion.

I mean 88 PCR reaction tubes for each condition and biorep, if 2.5 ug temple can be amplified per tube.

Julia, I do appreciate if you could help us to figure out the details regarding PCR.

From the context of your previous e-mail, I guess you mix 10 different Fwd primers to make a PCR mastermix.

Have you ever compared PCR efficacy between 10 mixed Fwd +Rev vs (one Fwd +Rev) X10?

For example,

Option1: 90 tubes with mixed 10 Fwd(1-10) +Rev1

Option2: 9 tubes with Fwd1+Rev1, 9 tubes with Fwd2+Rev1, 9 tubes with Fwd3+Rev1, ,,,,,,, , 9 tubes with Fwd10+Rev1 (=total 90 tubes) and then mix before running gel(??)

Which one is better?

The option 1 may be practically easier.

Best,

Yoshi

[athe...@gmail.com]

Thank you, Yoshi. You make the question more clear.

On Tuesday, November 7, 2017 at 7:02:53 PM UTC-8, Yoshi wrote:

[Julia Joung]

Hi Mei and Yoshi,

To clarify the protocol, for each of the forward primers, run 8-9 PCR reactions for 88 reactions total.

I have not done a direct comparison between Option 1 and 2 to determine which is more efficient, but I have tried both and have had good results from both. For 88 reactions, practically speaking it would be easier to make a 90x PCR mastermix with mixed 10 Forward primers (9x1.25uL of each 10uM Forward primer) and 1 reverse primer. 

Best,

Julia

--

You received this message because you are subscribed to the Google Groups "Genome Engineering using CRISPR/Cas Systems" group.

To unsubscribe from this group and stop receiving emails from it, send an email to crispr+unsubscribe@googlegroups.com.

[Yoshihiro Gocho]

Hi Julia,

Thank you so much for sharing your experience. I will go forward to a NGS part.

Best,

Yoshi

--
You received this message because you are subscribed to a topic in the Google Groups "Genome Engineering using CRISPR/Cas Systems" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/crispr/Qk4V_oy82iM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to crispr+unsubscribe@googlegroups.com.

[Yoshihiro Gocho]

We are still struggling with our PCR optimization.

I try to keep less cycle number of PCR to reduce potential biases introduced during amplification.

Is there any limitation of PCR cycle number for DNA extracted from harvested cells after selection?

23 PCR cycles are used in the Nature protocol. When we put 2.5ug of DNA into 50ul reaction, 23 cycles was not good enough to get the PCR products.

32 cycles with 2.5ug of DNA worked well. However, I have concerns about biases.

Should we optimize cycle number for each sample?

Thanks,

Yoshi

[Julia Joung]

Hi Yoshi,

I usually try to avoid going above 30 cycles to reduce potential bias. Did you try optimizing the amount of gDNA you are loading to see if you can improve the efficiency of your PCR?

Best,

Julia

[athe...@gmail.com]

Hi Julia

Thanks a lot for these discussions. I have some questions about the deep sequencing.

From your paper: you sequence the samples on Miseq or Hiseq, with 80 cycles of read1 and 8 cycles of index1. you recommend 5%phix for Miseq and 20% phix for Nextseq. you recommend aiming for a coverage of >100 reads per sgRNA in the library.

1. What are the output data, coverage and the total number of reads per gRNA library?  Is it correct to calculate in this way?

The total reads per library=10^5 ( number of sgRNA in the library) X100 (reading depth)=10 million/ library

The output data per library=88bp ( reads length) X10^5 ( number of sgRNA in the library) X100 (reading depth)=8.8X10^7

2. Our sequencing company will sequence the samples on Hiseq for us. They recommend 30% phix. Is this percentage appropriate?

I really appreciate your help.

Best,

Mei

On Tuesday, November 7, 2017 at 7:02:53 PM UTC-8, Yoshi wrote:

[Yoshihiro Gocho]

Hi Julia,

We optimized PCR condition (the amount of DNA and primer and PCR cycle).

A small amount of primer does not work for amplifying 2.5 micro gram of gDNA.

Our optimized condition is like this.

Template 2.5 ug

2X Hifi master mix 25 ul

Fwd primer 4.5ul (final conc: 0.9uM)

Rev primer 4.5ul (final conc: 0.9uM)

dd H2O other

Total 50ul
 

28 Cycles

Thanks,

Yoshi

[Julia Joung]

Hi Mei, Yoshi,

We recently found that PhiX is not necessary on the MiSeq and NextSeq if you use all 10 F primers (with staggered bases) to amplify your sgRNA library. We recommend aiming for >100 reads for the plasmid library QC and >500 reads for the screening library readout.

Given these parameters, the total reads per library would be 50 million reads/library, or 50 million reads/screening condition (i.e., control biorep 1) and the output data would be ~4.4 billion bases. I think you can probably get away with 0-10% PhiX on the HiSeq since I was able to get away with none on the NextSeq, but I haven't used the HiSeq in a while so check again with your sequencing company whether they would still recommend 30% PhiX on the HiSeq given that you do not need any for the NextSeq.

Thanks Yoshi for posting your optimized primer conditions for others to test - best of luck with your screening analysis!

Best,

Julia