search

tacazares / SeedMatchR

Off-target analysis of siRNA knock down paired with RNAseq
https://tacazares.github.io/SeedMatchR/
Other
6 stars 3 forks source link

Data interpretation #5

Open Domikohlh opened 3 months ago

Domikohlh commented 3 months ago

Hi,

I have a few questions about the software.

  1. I attempt to use siRNA-seq to generate the DGE result. Will different sequencing methods affect the data interpretation and accuracy?
  2. I am a bit confused about the data generated by de_fc_ecdf() and ecdf_stat_test(). Your paper mentioned that you would expect a distribution shift to the left for the seed-mediated effects, but how could it imply the off-target toxicity effects?
tacazares commented 3 months ago

Hello @Domikohlh, Sorry for the delay! This package is meant to be used with bulk RNA-seq data that has been generated from siRNA knockdown experiments. I am not familiar with siRNA-seq, but if you tell me a bit more I might be able to let you know if it will be compatible.

The shift to the left for siRNA target genes (or miRNA target genes) would imply that those genes are potentially targeted by the siRNA. The D statistic reported is the maximum difference between the background distribution and the target gene distribution. Our null hypothesis is that there should be no difference between the distribution of those sets of genes. The p-value is generated from the KS test.

Here are some documents that might help. Feel free to post more questions and I will try to help!

The literature has used ECDFs to gauge siRNA off target effects. This package and article cite the following example: https://academic.oup.com/nar/article/50/12/6656/6613918

ECDF overview: https://www.library.virginia.edu/data/articles/understanding-empirical-cumulative-distribution-functions This is a great package of different approaches for comparing ECDFS: https://twosampletest.com/reference/two_sample.html#ref-usage

Domikohlh commented 3 months ago

Hi @tacazares,

This is absolutely fine.

I’m sorry that I typed that wrong. That should be single-cell RNA sequencing, not siRNA-seq. I know scRNA-seq will have a slightly more info than the bulky one so it’s just to make sure that the gene expression profile generated by both single-cell and bulky would be the same.

I will take a look of the paper. Thanks a lot!

AR-Shicheng commented 1 month ago

Hi @tacazares,

This is absolutely fine.

I’m sorry that I typed that wrong. That should be single-cell RNA sequencing, not siRNA-seq. I know scRNA-seq will have a slightly more info than the bulky one so it’s just to make sure that the gene expression profile generated by both single-cell and bulky would be the same.

I will take a look of the paper. Thanks a lot!

It is a great idea to perform a scRNA-seq to siRNA knockdown samples. It will be exciting to see data public data and check what it looks like for siRNA knock-down assay.