Leandros Boukas
@LeandrosBoukas
Followers
315
Following
2K
Media
37
Statuses
798
Human Genetics, Epigenetics/Gene Regulation, Population Genetics. Medical Genetics & Genomics Fellow at @harvardmed and @BostonChildrens. MD, PhD
Joined April 2019
A patient derived missense mouse model of Kabuki syndrome 1
biorxiv.org
Kabuki syndrome (KS) is a rare cause of intellectual disability resulting from heterozygous pathogenic variants in the gene encoding the histone methyltransferase KMT2D. A previously established...
0
1
4
De novo status can often upgrade a variant from VUS to pathogenic, but for many families undergoing genetic testing it's not feasible to collect samples from both biological parents. Hoping duoNovo can help address this.
1
0
0
Looking forward to presenting duoNovo at ASHG this week. We use long-read sequencing to detect de novo variants without using both parents. It's conceptually straightforward, and performs very accurately among variants likely to be clinically relevant. https://t.co/zWrlauhAdq
medrxiv.org
While de novo variants cause many Mendelian disorders, their detection currently requires sequencing of the proband and both biological parents. This is not feasible when only one parent is availab...
1
0
2
We are excited to share GPN-Star, a cost-effective, biologically grounded genomic language modeling framework that achieves state-of-the-art performance across a wide range of variant effect prediction tasks relevant to human genetics. https://t.co/FTm3byYp67 (1/n)
17
161
518
The big one is finally out!! In this paper, we set out to provide insight into the fundamental question; How do the individual cells from complex tissues regulate their proteomes? Brief summary of our findings 👇 https://t.co/4dRXpGJTGn
biorxiv.org
Protein synthesis and clearance are major regulatory steps of gene expression, but their in vivo regulatory roles across the cells comprising complex tissues remains unexplored. Here, we systematic...
6
39
202
One of the greats of 20th century biology. What a life. https://t.co/0uduZialHU
0
0
0
New preprint on a surprising question - with a pangenome reference, *what is a genetic variant?* https://t.co/3CivztB04v With Pouria Salehi Nowbandani, Shenghan Zhang, Haoyang Hu, and Heng Li @lh3lh3
biorxiv.org
Structural variation causes some human haplotypes to align poorly with the linear reference genome, leading to ‘reference bias’. A pangenome reference graph could ameliorate this bias by relating a...
2
54
216
Creative use of transformers. Looking forward to diving into this
Negative binomial regression and inference using a pre-trained transformer -
0
0
0
New preprint! We solve a mystery you didn't know existed. Mitotic cells lack new transcription but require ongoing translation. Interphase mRNA half life is only 2-4 hrs. So how do cells arrest in mitosis for hours without depleting their transcriptomes? https://t.co/2CCq7AgYWS
biorxiv.org
In the presence of cell division errors, mammalian cells can pause in mitosis for tens of hours with little to no transcription, while still requiring continued translation for viability. These...
4
37
153
BTW this work was originally presented at ICML Workshop on Comp Bio in 2023. That’s 2 years ago!!! This kind of critical but fair view should not come to light this late, especially when so many gLMs are being released every week!
Our work on "Evaluating the representational power of pre-trained DNA language models for regulatory genomics" led by @AmberZqt with help from @NiraliSomia & @stevenyuyy is finally published in Genome Biology! Check it out! https://t.co/AFBC9Qu4x3
1
2
37
Not just semantics and words. It has deep implications for the models ability to generalize. It can be very accurate on certain types of test set constructions & learn a completely nonsensical underlying model, which will catastrophically fail on real world applications. 2/
1
6
61
Announcing Ambient Protein Diffusion, a state-of-the-art 17M-params generative model for protein structures. Diversity improves by 91% and designability by 26% over previous 200M SOTA model for long proteins. The trick? Treat low pLDDT AlphaFold predictions as low-quality data
4
42
192
New from the Rouhanifard Lab! What if we could sort single cells based on whether a specific gene was actively being transcribed—and then ask what drives that burst? Introducing: NuclampFISH
4
26
123
BLOG ALERT! 🧬 If you do RNA-seq, ATAC-seq, ChIP-seq, or modeling thereof, you may be overlooking LD score regression methods. These should be standard tools to study genes, variants, and regions en masse, but they are hard to understand. I wrote intros:
1
13
73
Really nice mini-blogs from @ekernf01 on the zoo of LD-score regression models for heritability and mediation analysis:
2
4
25
@LeandrosBoukas @doctorveera At the RET locus specific haplotypes of non coding SNPs can lead to severe loss of RET expression almost as much as heterozygous coding variant (50%). (PMID: 37948459) The odds ratio from these haplotypes is extremely high for Hirschsprung disease (PMID: 34782358)
1
2
2
you could follow the people i follow on twitter for several months catching tidbits about AI, or you could simply read lectures 3 and 5 from this brilliantly pithy lecture series by Radford Neal. https://t.co/YhRNgOF6Yf
1
1
1
If you’ve ever wondered about the significance of differences among mutational signature profiles, check out our new Aggregate Mutation Spectrum Distance (AMSD) preprint co-led by Sam Hart, @AlisonFeder with @nl_alcala
1
4
14
Plasmid-based reporter assays are the bedrock of regulatory genomics. But a basic question has gone unanswered for decades: Do chromatin architectures form on plasmids transfected into mammalian cells—and does it matter? We finally have answers. https://t.co/34hcyRPN0z
biorxiv.org
Plasmids have fundamentally transformed how we resolve regulatory grammar across the tree of life. However, although chromatin plays an integral role in regulating the function of regulatory elements...
6
84
308