Krishnaswamylab.org

HiPoNet: A Multi-View Simplicial Complex Network for High Dimensional Point-Cloud and Single-Cell Data In this paper, we propose HiPoNet, an end-to-end differentiable neural network for regression, classification, and representation learning on high-dimensional point clouds. Our work is motivated by si...

(10/N) If you are interested check out our preprint and code! Thanks for reading!
Arxiv: arxiv.org/abs/2502.07746
Code: github.com/Krishnaswamy...

(9/N) Grateful to lead authors Siddharth Viswanath and Hiren Madhu, and coauthors Dhananjay Bhaskar, Jake Kovalic, Dave Johnson, Chris Tape, Ian Adelstein, Rex Ying and Michael Perlmutter!

(8/N) ✨ In short, HiPoNet is a neural network that takes in a whole point cloud, and utilizes methods from geometry and topology to derive features of the point cloud.
→ Multiple learned views
→ Simplicial wavelets to extract multi-scale structure

(7/N) What’s more, HiPoNet’s learned feature weights are interpretable. For example, immune-related markers like CD11b, CD118, and FOXP3 consistently emerge as important across views, aligning with known tumor-immune biology.

(6/N) We find that HiPONet consistently outperforms other methods on these datasets.

(5/N) We then evaluate HiPOnet on patient level classification tasks using single cell and spatial proteomics data:

(4/N) We first check that HiPoNet preserves topological features of datasets by predicting persistence from the datasets:

(3/N) HiPoNet models each cloud as a simplicial complex, and uses simplicial wavelet transforms to capture point-cloud level embeddings. Not only that---it uses multiview learning, i.e. different projections of the features to capture different underlying factors in the data.

(2/N) Modern technologies such as mass cytometry or scRNA-seq now allow large cohorts of patients to be measured. Each patient is actually a large single cell dataset creating a high-dimensional point cloud. But these point clouds are far more complex than 3D shapes that ML methods have handled..

(1/N) Thrilled to share that our paper HiPoNet (High dimensional Point cloud Network) to be presented at NeurIPS 2025! HiPoNet treats an entire high-dimensional point cloud as a datapoint! It captures multi-scale geometry and topology of the cloud perform classification and regression tasks.

Good morning from Vancouver! 🌞⛰️ We’re at #ICML2025 presenting at GenBio + MAS. Come chat with us about machine learning in biology!

With @RobertXiangruTang, @ChenLiu, @DanqiLiao. Views like this make the science even sweeter.🧬🧠

#ML #ComputationalBiology #ICML #Vancouver

Defining and benchmarking open problems in single-cell analysis - Nature Biotechnology Nature Biotechnology - Defining and benchmarking open problems in single-cell analysis

Defining and benchmarking open problems in single-cell analysis - @krishnaswamylab.bsky.social @fabiantheis.bsky.social bit.ly/4l8dsiF

Flyer for the Kavli Awards Research in Progress of June 23 showing pictures of the speakers in the Kavli logo, with a schedule: 12:00 pm Dhananjay Bhaskar, PhD 2024 Kavli Postdoctoral Fellowship Krishnaswamy & Colón-Ramos labs 12:20 pm LaShae Nicholson, PhD 2022 Kavli Award for Academic Diversity Strittmatter lab 12:40 pm José Jaime Martínez-Magaña, PhD 2023 Kavli Award for Academic Diversity Montalvo-Ortiz lab 1:00 pm Kevin Chen, PhD 2024 Kavli Postdoctoral Fellowship Emonet & Clark labs 1:20 pm break 1:30 pm Aaron Kuan, PhD & Joerg Bewersdorf, PhD 2024 Kavli Teams Award 1:50 pm Enock Teefe, MD 2023 Kavli Award for Academic Diversity Pittenger & Fernandez labs 2:10 pm Hongyan Hao, PhD 2024 Kavli Postdoctoral Fellowship Hammarlund & De Camilli labs

Mark your calendars 🗓️ for June 23 for a Kavli Awards Research in Progress!🧠

Hear from Drs. Bhaskar (@krishnaswamylab.bsky.social‬), @lashaeneuroxp.bsky.social‬, Chen (@emonetlab.bsky.social)‬, Martinez-Magaña (@janitzamontalvo.bsky.social), @kuanawanda.bsky.social‬, Teefe, & ‪‪@haohy.bsky.social‬.

Accelerated learning of a noninvasive human brain-computer interface via manifold geometry Brain-computer interfaces (BCIs) promise to restore and enhance a wide range of human capabilities. However, a barrier to the adoption of BCIs is how long it can take users to learn to control them. W...

New preprint! Excited to share our latest work “Accelerated learning of a noninvasive human brain-computer interface via manifold geometry” ft. outstanding former undergraduate Chandra Fincke, @glajoie.bsky.social, @krishnaswamylab.bsky.social, and @wutsaiyale.bsky.social's Nick Turk-Browne 1/8

Excited to organize this! Should be an exciting event!!

Great ideas to introduce to the community!! Enjoyed the talk!

Join us for an exciting session in @jointmath.bsky.social tomorrow! Here is our revised schedule!

We are at @jointmath.bsky.social Jan 8-11!
1/8: My talk at Topological Machine Learning I
1/8: Dhananjay Bhaskar at Contributed Paper session on Topology
1/11 join our special session on Mathematical and Computational Oncology
Also check out @pseudomanifold.topology.rocks on 1/9!!

Merry Xmas!

Join the Krishnaswamy Lab We work on developing foundational mathematical <span class=

Looking for talented postdocs in deep learning for neuroscience and neuromodulation with the Murat Gunel lab in neurosurgery (medicine.yale.edu/lab/gunel/)! Projects include brain signal decoding, dynamics modeling for neuromodulation ASD and other conditions. See: krishnaswamylab.org/join

Frohe Weinachten!

GitHub - KrishnaswamyLab/GSPA-manuscript-analyses: Learning meaningful representations of genes Learning meaningful representations of genes. Contribute to KrishnaswamyLab/GSPA-manuscript-analyses development by creating an account on GitHub.

(10/n) For code see github.com/Krishnaswamy.... Thanks for reading!

(9/n) GSPA-Pt can be used to map patient sample manifolds. We mapped 48 melanoma patient scRNA-seq samples and classified response from the patient embedding using logistic regression. The GSPA-Pt gene embeddings achieved the highest classification performance.

(8/n) GSPA-LR concatenates ligand (L) and receptor (R) for a pair representation. LR pair modules captures a diverse range of LR profiles, finer the cluster-level analysis. For example, in skin cells, Module 5 includes Ccl5–Ccr5 link present in AG and AG CPI, epithelial, myeloid and T cells.

(7/n) With the Nik Joshi Lab at Yale we presented a new dataset of 39K CD8+ cells from LCMV infections. Interestingly only GSPA-based gene localization finds genes associated with type 1 interferon signaling. DE requires clustering, but DE genes in clusters do not reveal this signature!

(6/n) GSPA facilitates a measure called gene localization, via comparison to a uniform distribution. For PBMCs, cell embeddings using all genes versus only top localized genes showed similar structure, suggesting that localized genes capture cell-type variation and geometry.

(5/n) We show that for (PBMCs), GSPA with compressed wavelets grouped cell-type specific genes from PanglaoDB. In embryoid-body lineages early trajectory genes linked to embryonic stem cells (NANOG, POU5F1) and late trajectory genes linked to hemangioblast-specific signatures (CD34, PECAM1, TAL1).

(4/n) Rather than transposing the cellxgene matrix for gene embeddings, we frame genes as signals on a cell–cell graph, decomposed using graph-based diffusion wavelets. Diffusion wavelets use differences of powers of P. Then we then reduce dimensionality with an autoencoder.

(3/n) Low dimensional embeddings like UMAP, PHATE, tSNE have been vital in understanding geometry and subpopulations from single cell data. Gene expression is also highly organized (pathways etc). But due to noise, directly apply these embeddings to genes has not worked.

(2/n)This work is lead by Aarthi Venkat, Sam Leone. We couldn't have done it without our collaborators Scott Youlton, Eric Fagerberg, John Attanasio from Nik Joshi Lab at Yale and Michael Perlmutter from Boise State University!