Now out at #jbiolchem www.sciencedirect.com/science/arti...

Our full-day mini AI hackathon on genomic language models starts soon: 8.30am-8.30pm! With a whole team, we aim to test how gLMs work for two tasks:
1. SNP2GEX: personal variants to gene expression (unseen individuals & genes);
2. Seq2CellxTF: short regulatory to TF binding (unseen cell & TFs).

Open source @github.com github.com/jasonwong-la.... And thank you @yuanhuahuang.bsky.social for the amazing collaboration!

Great to finally see LocusMasterTE in print by PhD student, Sojung Lee. We developed a bioinformatics method to leverage long read sequencing data to improve expression quantification of transposable elements genomebiology.biomedcentral.com/articles/10.... #GenomeBiology

A case where the oodles of public lung cell line data generated as a result of COVID can be put to good use in cancer! Well done to Sojung Lee (PhD student), for the massive effort in data mining. (9/n)

Given that ERVK-7 expression is a potential biomarker for ICB response in LUAD, ERVK-7.long may be a more specific and easier to quantify. Controlling the epigenetic regulation of ERVK-7.long may also alter the immune microenvironment of LUAD. (8/n)

In addition to cell type specificity, the promoter of ERVK-7.long is driven by NF-κB, likely via TNF-α, while ERVK-7.short is already known to be regulated by IRFs. (7/n)

The promoter of ERVK-7.long was found to be lineage-specific, with H3K4me3 deposition in LUAD cells. ScRNA-seq confirmed that ERVK-7.long expression is restricted largely to LUAD cells, fibroblasts, and AT-2 cells, which may be the cell type of origin of LUAD cells. (6/n)

ERVK-7.long is highly expressed in LUAD, contributing significantly to the overall expression of ERVK-7. This suggests that the epigenetic regulation of ERVK-7.long may be key to ERVK-7 overexpression in patients. (5/n)

HERVs and retroviruses are transcribed from the 5'LTR. Yet PacBio Iso-seq revealed alternative transcripts for ERVK-7 arising from upstream promoters, with the canonical TSS of the 5'LTR heavily methylated in A549 (ONT WGS) along with TFBS mutations. (4/n)

A key motivation for us is to understand the epigenetic regulation of ERVK-7 as it is frequently over-expressed but only amplified in ~10% of LUAD patients. (3/n)

HERVs are generally silent or dysfunctional in the human genome. Yet recent studies have shown that the expression of certain copies, such as ERVK-7 (HERV-K102), is associated with autoantibodies and immunotherapy response in lung cancer patients t.co/21yyHijgMd. (2/n)

First post on Bluesky 🦋. Happy to share some new work on the epigenetic regulation of a human endogenous retrovirus (HERV) in cancer. @biorxivpreprint.bsky.social bsky.app/profile/bior... 🧵 (1/n)