Charlie Wright

Computational/Chemical biologist working in cancer research @stjude.

Developing tools using #machinelearning & #highthroughput technology.

[Charlie Wright @biowright]

Check out this interesting work from the Stacey Ogden lab @StJudeResearch, out now in @NatureComms :

A case where virtual screening and MD simulations successfully predict binding residues and allosteric sites of SMO. This was a great collaboration!

https://t.co/9xMcBqY3bQ

[Huang_Nanopore @Huang_Nanopore]

New Huang lab publication at Nature. Our newly developed technique: transient Pore Analyte Looping (tPAL) now achieves “chop and measure” nanopore sequencing of peptide, in an amino acid by amino acid manner. https://t.co/YMUPMckpnx

[Charlie Wright @biowright]

@StJudeResearch @ADurbinLab @hatleylab @mikedyerlab @stjudegraduate @AACR Congratulations to all the presenters!

[Charlie Wright @biowright]

@rust_ruslan Thanks

[St. Jude Research @StJudeResearch]

St. Jude researchers developed Combocat, a machine learning-powered platform that rapidly screens drug combinations at scale. In a proof-of-principle test, 9,045 drug pairs were screened, uncovering new synergistic therapies. https://t.co/2KXZXX0yCS

[Charlie Wright @biowright]

I once heard a talk from @JEFworks explaining how she uses Moran's I to measure spatial autocorrelation in spatial transcriptomics. I wondered if this might be applied to these 9000+ synergy matrices. It turned out to be a valuable filtering metric!

[Charlie Wright @biowright]

To scale higher, we developed the "sparse mode" of Combocat, which introduces several tricks to dramatically boost throughput...like use of 1536-well plates, and generating 'sparse' matrices.

[Charlie Wright @biowright]

Beyond throughput, Combocat emphasizes data quality and interpretability. Various QC and filtering options are used to enable direct comparisons of results across experiments.

[Charlie Wright @biowright]

Lastly, Combocat is fully open-source. We provide code, documentation, and example workflows at
https://t.co/eFDfYGyEZW

Have an interest in drug combinations? Reach out!

[Charlie Wright @biowright]

Dose density is important. Combocat's "dense mode" screens combinations in a 10×10 dose format - in triplicate.

We tested > 800 combos this way in various cell models like cancer, 'normal', and bacteria.

[Charlie Wright @biowright]

Drug combinations are central to modern medicine, but their discovery is still highly limited by scale, cost, and analytical complexity.

Combocat addresses this by integrating experimental design, automation-friendly layouts, and standardized analysis into a single framework.

[Charlie Wright @biowright]

Excited to introduce Combocat, an open-source platform for large-scale drug combination screening and analysis published today in @NatureComms :

https://t.co/QJDyu1LjW4

Brief overview..🧵

[Charlie Wright @biowright]

This year at the Bringing Chemistry to Medicine symposium, I got to discuss how drug combination discovery can be accelerated.
Had several great discussions while presenting this poster, which spans the fields of computational and chemical biology. A very inspiring symposium!

[Charlie Wright @biowright]

Had the incredible opportunity to lead a team in the St. Jude #KIDS25 Biohackathon.
Over just 3 days, we developed machine learning approaches to predict cancer dependencies from methylation data. Everyone put in a ton of effort, and we all learned a lot!💻

[Charlie Wright @biowright]

@Larsen_in_Lab @HannaLabPurdue Congrats!

[Charlie Wright @biowright]

@mfranz_on @threejs Neat! Is it possible to model well-described attractors like Lorenz?

[Charlie Wright @biowright]

@ItaiYanai If only..

[Charlie Wright @biowright]

@RConsortium Julia's instructional videos are pure gold!