Barkai Lab

How do transcription factors (TFs) that recognize the same DNA motif end up binding different places in the genome? 🧬 We're excited to show you how we tackle this question in our new bioRxiv pre-print! doi.org/10.64898/2026.…

Our work suggests that TF specificity isn’t just encoded in DNA motifs or chromatin state... Intrinsic TF features—especially disordered non-DBD regions—play a central role in determining where it binds, highlighting mechanisms beyond motif or chromatin-based models.

So what’s driving this genomic selectivity? Our data point to disordered regions outside the DNA-binding domain as major determinants of binding strength and genomic preference, even in the absence of specific cofactors.

(2/8) By @JingLiu114 guided by @FJonas15 @NaamaBarkai authors.elsevier.com/c/1mdGb3vVUPZN…

(1/8) Still wondering how low complexity intrinsically disordered regions (IDRs) guide transcription factors (TFs) to bind in genomes? Previous studies on sequence grammars; our new @MolCell paper takes the next step: de novo design of functional TF IDRs! cell.com/molecular-cell…

(8/8) Huge thanks to the whole Barkai lab for their support and the reviewers for their constructive feedback! And also, happy Chinese New Year! 🐎🎆

(7/8) We’ve moved from observing and perturbing TF IDRs to engineering them from de novo. We’re excited to exploit these grammars for synthetic biology and to see if these "simple rules" hold true across other species and protein classes! 🚀

A preprint‼️that's bound to ruffle some 🪶 "Widespread DNA off-targeting confounds studies of RNA chromatin occupancy" led by our @MicahGoldrich and Louis Delhaye from @pieter_mestdagh. TL;DR we show that many of lncRNA chromatin occupancy maps are flawed🧵biorxiv.org/content/10.110…

Our analysis of the data suggests that there may not be a universal answer. Rather, we favor cross-validating the conclusions across definitions (e.g, thresholds, peak definition), and against orthogonal biological datasets.

There are pros and cons to focusing on top-bound targets vs. including all significant peaks. Can we assume that binding signals are proportional to binding affinities? And what would be the threshold for 'top-bound'?

A recent study reported a surprisingly low overlap between TF binding and regulatory targets. But this may depend on how binding sites are defined, as we report; Should one include all statistically significant peaks or focus on top-bound targets? biorxiv.org/content/10.110…