1
$\begingroup$

I have some eclipseq data, three IP and three input datasets, and have the alignments for all three. I used Bedtools "Intersect Interval" with the genome gtf file (hg19) I used to align them with, and that creates a bed file with gene names for each peak. I then wrote an R script to sum the clipper score for each gene within a sample, then do a simple student's t test across groups. I now have a gene list and p values for differential expression, but not sure if this is correct.

I've tried looking at a lot of sources but cant really find any canonical way to generate a gene list from bed files.

I've also tried IDR and it doesn't work on the clipper output, unfortunately.

Thanks!

Edit: eCLIP-sequencing, it's essentially a variant of iCLIP-sequencing and data analysis methods are transferrable, to my knowledge.

I have alignments (bam files) made with Star aligner against Hg19. After deduplication and sorting, I called peaks using Clipper (a peak-caller made by the creators of eCLIP). It outputs bed files with scores for each sample.

My question is how exactly do I turn this bed file with scores into a gene list? Is my method above correct?

$\endgroup$
3
  • $\begingroup$ Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. $\endgroup$
    – Community Bot
    Dec 31, 2023 at 7:14
  • $\begingroup$ What is "eclipseq data"? Can you please provide some example input files? Anything you can add to clarify the most important problem you're having will be useful for people who are answering this. $\endgroup$
    – gringer
    Dec 31, 2023 at 8:46
  • $\begingroup$ Added more information. Thanks! $\endgroup$
    – user18763
    Dec 31, 2023 at 18:27

2 Answers 2

1
$\begingroup$

As someone who works with eCLIP data from ENCODE, I think I can help out.

You need to think of this more similar to ATAC or RNA-Seq, whereby you actually get a counts matrix, normalize that, and then apply math.

Another good idea would be to look through this eCLIP analysis document provided by Gene Yeo of ENCODE for more specific ideas.

https://www.encodeproject.org/documents/1f171ac6-a36a-41ac-b632-741aeb47aad2/@@download/attachment/eCLIP_analysisSOP_v2.3.pdf

$\endgroup$
0
$\begingroup$

As others have said it's not very clear what you're asking here, but I'm assuming you want to know how to determine relevant genes that were enriched in your IP.

As per the group behind eCLIP, Van Nostrand et al. 2020 : "standard peak analysis used the set of peaks identified as irreproducible discovery rate (IDR) reproducible and meeting fold-enrichment (≥ 8-fold) and significance (p value ≤10−3) in immunoprecipitation versus paired size-matched input."

If those cutoffs are too stringent you could change them to fit the context of your project. And what do you mean by differential expression? It sounds like you're addressing binding.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.