Program/Tool that can determine whether two profiles are likely to be from the same cell line, based on sequence analysis (RNA seq and Chip Seq).

Question

[David_emir]

Hello alI,

I am in search of a program/tool which can take the given RNA-seq or ChIP-seq data, what program can be used to determine whether two profiles are likely to be from the same cell line, based on sequence analysis. Preferably this should be a standalone program that runs at the command line, is passed two aligned sequence (BAM) files, ideally is built on publicly available tools, and completes in less than 5 minutes. Please let me know your thoughts on this. Thanks a lot for your help.

[dariober]

I don't think it's even theoretically possible to say whether two samples come from the same cell line. With just two samples there is no way to tell how much of the difference is due to the cell line and how much is due to technical (ir)reproducibility or else.

By the way, requiring to complete in less than 5 minutes it's probably going to be tough.

[SP]

I agree with @dariober, I don't think so there is any tool. But if you really want to compare two ChIP-seq profiles are from single cell-line you should probably investigate cell type specific binding pattern (peaks) of particular factor. You might need to set a threshold for similarity of binding profile but this work might take some time to think the parameters you want to include. I believe you have to write code by yourself and time would be a big constrain. All the best

Answer 1

[Sukhi Singh]

As the other guys have mentioned already, this could work only to an extent.

Starting for raw files and even before that to the real experimentation, the samples are susceptible to variability at many places, eg

• contamination,
• coherence of technical/biological ir/reproducibility,
• %GC bias (machine and PCR dependent),
• mapper's non coherence (gaps, mismatches) etc.

Still you could achieve the answer if the files you are comparing, are very dissimilar but replicates/single conditions have similar profiles.

If you go bit downstream, you use tools like bamCompare and bamMatcher

Also, check this post.