Mesude Bicak (@MesudeBicak) header
Mesude Bicak (@MesudeBicak) profile picture
Manhattan, NY
Joined September 2011
Photo by Mesude Bicak (@MesudeBicak)Photo by Mesude Bicak (@MesudeBicak)Photo by Mesude Bicak (@MesudeBicak)Photo by Mesude Bicak (@MesudeBicak)Photo by Mesude Bicak (@MesudeBicak)Photo by Mesude Bicak (@MesudeBicak)
Honored to receive Mike Slive Foundation award for our efforts towards an off-the-shelf prostate cancer vaccine. Special thanks to my rockstar colleagues for laying all the pioneering innovative foundation @timodonnell @JuliaKodysh @CimenBozkus_C @Mansi_Saxenm02 @BhardwajLab
2
1
11
1,632
🗣️ “I was the first person ever at a professional event to fly the flag of North Cyprus.” Professional boxer Sugar Atahasan recalls carrying the flag of the Turkish Republic of Northern Cyprus into the ring. UNSEEN: The Athletes of Northern Cyprus The full documentary is out on 9 November on the Freedom and Fairness for Northern Cyprus YouTube channel.
1
4
11
476
Mesude Bicak retweeted
FYI I wrapped a lot of immuno-informatics predictors so you don't have to: openvax.github.io/mhctools/ (eg NetMHCpan, MHCflurry, Pepsickle, DeepTAP, ERAMER, BigMHC, PRIME, &c)
2
25
1
152
8,221
Mesude Bicak retweeted
MHCflurry 2.3 is out Lots of internal improvements to make both training and prediction faster on multi-GPU Also trained a slightly better model (first new weights in a long time) Thanks to @nvidia / @AllThingsApx for the Brev credits Try it out and tell me what I broke
3
10
1
70
4,843
Mesude Bicak retweeted
Check out our latest review👏! We dive deep into how cDC1s orchestrate anti-tumor immunity through cross-presentation, immune networking, and spatial organization.
Ward, Kar, Balan, and Bhardwaj @BhardwajLab review how conventional type 1 #DendriticCells (cDC1s) coordinate antitumor immunity through cross-presentation, immune cell networking, and spatial organization in tumors. hubs.la/Q04kVmqr0 #CancerFocus #Immunotherapy
11
41
4,941
Mesude Bicak retweeted
MHCflurry 2.3.0rc1 is on PyPI! It’s a very big change set that moves as much work as possible out of NumPy / Pandas and into PyTorch. It also adds multi-GPU support for both training and inference. Training time on a 8xA100 machine is about an hour (down from a week)
1
3
27
2,057
Mesude Bicak retweeted
Guillaume’s latest paper went online yesterday @CellRepMed! This research shows that a quad-blockade overcomes immune resistance in colorectal cancer by reprogramming macrophage-T cell crosstalk to induce high rates of tumor clearance. Check it out here: cell.com/cell-reports-medici…
2
5
14
1,088
Jonathan Anker, MD, PhD, has been appointed as Instructor in Hem/Onc. He provides clinical services focused on genitourinary medical oncology and conducts research on tumor immunology and immunotherapy for urothelial cancer. @JonAnker1 @IcahnMountSinai scholars.mssm.edu/en/persons…
3
4
1
33
2,235
From @BhardwajLab and colleagues: Harnessing Notch Signaling to Enhance the Generation and Functionality of Human Conventional Type I Dendritic Cells for Cancer Immunotherapy Applications pubmed.ncbi.nlm.nih.gov/4060… mountsinai.org/about/newsroo… @IcahnMountSinai #cancervaccine
4
12
677
Just published in Cancer Immunology Research! Our lab reports the first scalable, serum-free production of bona fide, functional human cDC1s—generating ~3 billion from 1M HSCs. Read more: mountsinai.org/about/newsroo…
1
5
26
1,196
Atezolizumab in combination with personalized neoantigen vaccine for urothelial cancer. urotoday.com/recent-abstract… via @urotoday @BhardwajLab @MattGalsky @IcahnMountSinai #bladdercancer
2
5
529
Nina Bhardwaj, MD, PhD, is pleased to serve as Co-Chair of #CICON25. cancerimmunotherapyconferenc… Registration is open now. @BhardwajLab
Join us at #CICON25, from September 10-12, as we explore exciting new research in cancer immunotherapy, and shape new strategies that translate scientific discoveries into life-changing outcomes! Review the program, and register today: bit.ly/4cjiDJ4 #CancerResearch
5
9
1,299
Matt Galsky, Nina Bhardwaj, and team show that custom-made vaccine, in combination with immune checkpoint inhibitor, sparks strong immune responses in people with #BladderCancer. mountsinai.org/about/newsroo… @MattGalsky @BhardwajLab @IcahnMountSinai
3
1
1
5
2,359
I can't believe MHCflurry is still crushing, it also is pretty much the only thing I can use to in any way enrich for hits in targeted MS immunopeptidomics. @timodonnell did a great job dragging a prototype into something that's extremely useful many years later.
Comprehensive evaluation and interpretative insights of peptide-HLA binding prediction tools using explainable artificial intelligence 1. This study presents the most extensive benchmarking to date of peptide–HLA-I binding prediction tools, evaluating 17 models on over 290,000 peptides across 44 HLA-I alleles, using a rigorously curated, independent dataset. 2. Self-attention-based models STMHCpan and BigMHC ranked highest in performance, with STMHCpan achieving AUROC 0.961 and AUPRC 0.965—outperforming all other methods across nearly all peptide lengths and HLA variants. 3. Models trained on eluted ligand (EL) data outperformed those trained on binding affinity (BA) data. NetMHCpan4.1_EL and MHCflurry2.0_PS both surpassed their BA-trained counterparts, demonstrating the importance of high-quality MS-derived datasets. 4. Capsule network model CapsNet-MHC_AN showed strong performance and low computational cost, suggesting spatial feature hierarchies offer a viable alternative to transformer-based architectures for peptide–HLA binding prediction. 5. Explainability analysis using SHAP and LIME revealed critical binding positions—especially P2 and P9—across alleles. SHAP provided more stable and biologically consistent interpretations, particularly for rare alleles like HLA-C*05:01. 6. Computational alanine scanning validated model explanations by comparing attribution scores with predicted ΔΔG values, confirming SHAP's ability to highlight residues that contribute most to binding energy. 7. STMHCpan, BigMHC, and MHCflurry2.0_PS consistently identified key binding residues in test peptides. Their predictions aligned well with energy-based structural models, while tools with low attribution values showed higher false negative rates. 8. 9-mer peptides were best predicted across all tools. HLA-A and HLA-B alleles had better predictive accuracy than HLA-C, largely due to greater training data availability—underscoring the need to expand training sets for underrepresented alleles. 9. Ensemble models combining multiple predictors, such as MHCflurry2.0’s logistic integration of AP and BA modules, delivered improved performance over standalone components, pointing to the power of hybrid approaches. 10. This study not only ranks current peptide–HLA-I predictors but also introduces an open benchmarking platform with interpretable AI integration, offering a roadmap for the next generation of immunoinformatics tools. 💻Code: github.com/YY-TMU/MHCI_bench… 📜Paper: biorxiv.org/content/10.1101/… #HLA #peptidebinding #immunoinformatics #bioinformatics #machinelearning #SHAP #LIME #benchmarking #AI4Science #neoantigenprediction
2
9
33
3,630
Congratulations to several members of our Vaccine and Cell Therapy Laboratory, past and present, on the publication of our PGV neoantigen vaccine trial data in Cancer Discovery! 🥳
PGV001, a multi-peptide personalized neoantigen vaccine platform: Phase I study in patients with solid and hematological malignancies in the adjuvant setting aacrjournals.org/cancerdisco… @BhardwajLab Thomas Marron, MD, PhD @IcahnMountSinai
1
8
32
2,787
PGV001, a multi-peptide personalized neoantigen vaccine platform: Phase I study in patients with solid and hematological malignancies in the adjuvant setting aacrjournals.org/cancerdisco… @BhardwajLab Thomas Marron, MD, PhD @IcahnMountSinai
7
2
26
6,910
Congratulations to Nina Bhardwaj @BhardwajLab @IcahnMountSinai, named a Fellow of the @AACR Academy Class of 2025. aacr.org/about-the-aacr/news…
👏 Congratulations to CRI Scientific Advisory Council Associate Director Dr. Bhardwaj @BhardwajLab @IcahnMountSinai and CRI SAC member Dr. Vonderheide @PennMedicine, who have been named Fellows of the @AACR Academy Class of 2025! Learn more: bit.ly/3DuLRYy #AACRFellows
2
1
6
3,255
Congratulations @NinaBhardwaj !
👏 Congratulations to CRI Scientific Advisory Council Associate Director Dr. Bhardwaj @BhardwajLab @IcahnMountSinai and CRI SAC member Dr. Vonderheide @PennMedicine, who have been named Fellows of the @AACR Academy Class of 2025! Learn more: bit.ly/3DuLRYy #AACRFellows
2
15
856