Exporting count matrix

[hyunhwan-jeong]

Hello,

I am wondering how I can export the count matrix which is derived from the input parameters. I just only able to have the values of screening performance of the given simulated study as the output of Crispulator. Are you providing any output/function which reveals the count matrix of each sample? If it is not then could you advise me to implement the function?

Thank you,

[tlnagy]

Could you clarify what you mean by the count matrix?

[hyunhwan-jeong]

Oh sorry, I would explain it clearly. Crispulator simulates the deep sequencing, and we can have read numbers of each sgRNA for low/high reporter bin. I want to have the read numbers once I run the tool, and I called it count matrix because we can have the sgRNA read numbers for each bin with the form of n x 2 matrix (n is the number of sgRNA, and columns are for low/high bin). I hope I clarify it well.

Thank you

[tlnagy]

Sorry for not getting back to you sooner. Crispulator does generate your count matrix, but doesn't export it currently. It should be relatively easy to add.

[hyunhwan-jeong]

Hello @tlnagy,

I recently realized that you provide the bc_counts data frame, and it stores counts of each sgRNA from the high-throughput sequencing. Therefore, I think you don't have to add any additional function or script. Please correct me if I am wrong or close this issue if it is otherwise. I thank you for your attention to this request and I congratulate your recent publication of this great tool in the journal!

Thank you,

[tlnagy]

Let me know if bc_counts worked for you! If not, please open a new issue. I'm going to add an example to the docs based on your example in #52. I think it would help others in the future!

[hyunhwan-jeong]

I think it is a kind of okay if I want to build single counts for each group. However, I'd like to have any idea how to build the matrix under below conditions:

• Generate another bc_counts of un/pre-sorted cells with keeping counts of the high and low.
• Generate multiple replicates under the same configuration, I did it before, but not 100% sure my method is correct.

Thank you,

Hyun-Hwan Jeong

[tlnagy]

Generate another bc_counts of un/pre-sorted cells with keeping counts of the high and low.

Are you simulating a FACS Screen?

[hyunhwan-jeong]

Are you simulating a FACS Screen?

Yes, for the case.

[tlnagy]

Generate another bc_counts of un/pre-sorted cells with keeping counts of the high and low.

I realized that obs_phenotype was supposed to serve this purpose so I reverted the previous commit that deleted it. I fleshed out the docs further. http://tamasnagy.com/Crispulator.jl/latest/custom.html#Performing-the-screen-1 should help you with getting the observed phenotype, i.e. the phenotype on which the cells were sorted with FACS

EDIT: It sounds like you obs_phenotype for each individual cell, not on a per-guide level basis, correct?

I think that would involve exporting observed:

Crispulator.jl/src/simulation/selection.jl

Lines 20 to 22 in 46a804c

	observed = zeros(n_cells)
	@inbounds for i in 1:n_cells
	observed[i] = rand(Normal(cell_phenotypes[i], σ))

[hyunhwan-jeong]

Hello @tlnagy,
Thanks for your reply, but I do not want to have cell count, and my language made you confused. My question must be corrected as "Can we have read counts for sgRNA for [0,100%] bin or [25%,75%] bin?". I wanted to simulate a CRISPRn experiment like 1. In Figure 1 of the [1], you can see they collected and sequenced for "unsorted cells". Hope it makes you clear now.

Thank you and Best,

Hyun-Hwan Jeong

[tlnagy]

I see what you want. So I've had to rearchitect differences_between_bins to support this approach. It's almost ready. I'll push a PR once I'm done.

[hyunhwan-jeong]

I appreciate your support!

Kind Regards,

Hyun-Hwan Jeong

[tlnagy]

Hey @hyunhwaj, if you get a chance can you check out the notebook in #53? It has the Unsorted, GFP_low, GFP_high bins that you were looking for. Let me know over there if it works for you.

You should be able to check it out using

git fetch origin pull/53/head:add-multibin-support
git checkout add-multibin-support

in the Crispulator directory

[tlnagy]

I merged #53 into master so now you should be able to run

Pkg.checkout("Crispulator")

to give it a spin. Please open a new issue if you run into any issues or have suggestions!

[hyunhwan-jeong]

That's actually what I've wanted! I checked the notebook, and this ran without any problem. I truly appreciate your work.

Hyun-Hwan Jeong

[tlnagy]

Thanks! Let me know if you need anything else.