differential TFBS in trios

[Fenrir Sivar]

I'm running an analysis on families with rare diseases, in particular trios of mother/father/child or siblings.

What would be a good way to get some insight into which binding sites are affected in our target genelists between the probands and unaffected family members? I have WGS variant data of all my probands, siblings and parents.

I'm looking in particular for some example scripts or literature with methods in either R or python (anything else will do as well)

Best, leo

[Jaime Castro-Mondragon]

Hi Leo

One possibility is to intersect each of your sets of mutations with the predicted TFBSs in JASPAR, or Unibind (https://unibind.uio.no/  These are TFBSs predicted within ChIP-seq peaks using the Jaspar matrices)

You can use either bedtools, pyranges (python) or genomicranges (R/Bioconductor)

Good luck

Jaime

[Oriol Fornés]

Hi Leo,

MANTA2 was intended for precisely this use and can be queried both programatically and through a dedicated web server. It takes a VCF file as input and outputs the predicted impact of variants on TF binding sites. As Jaime suggested, then you could simply intersect the predictions from the proband with those from the parents/siblings.

Alternatively, you might want to look into DeepSEA, which also takes as input a VCF file and outputs the predicted impact of variants on a variety of genomics features including TFs.

[Fenrir Sivar]

Thank you both for the suggestions. It looks like I will have to build a local version for MANTA2, since the patient data should stay in a controlled environment. I'm familiar with MongoDB, Python and Flask so that should be doable.

This brings me to the following question: what are reasonable specs for the underlying server?

Is there a 2020 version for the data dumps available?

Best Leo

Op woensdag 7 oktober 2020 om 21:16:11 UTC+2 schreef oriol....@gmail.com:

[Oriol Fornés]

Hi Leo,

MANTA2 has not been updated since its inception in 2018, as it depends on TF binding sites predicted within ChIP-seq regions (i.e. similar to UniBind but it includes the impact of variants on TF binding).

Instructions for its installation can be found here (it should be a relatively easy exercise):

https://github.com/wassermanlab/MANTA2/blob/master/README.md#installation