TMAP2

Tree-based visualization for high-dimensional data. Organizes similar items into interactive tree structures. Ideal for chemical space, protein embeddings, single-cell data, or any high-dimensional dataset.

Why Trees?

Most dimensionality reduction tools (UMAP, t-SNE, PCA) produce point clouds. TMAP produces a tree, a connected structure where every point is linked to its neighbors through branches. This makes the layout itself explorable because you can follow branches, trace paths between any two points, and discover how regions connect.

For example, in a TMAP of pet breed images, following the branch from terriers toward cats reveals that the bridge between the two groups runs through chihuahuas and sphynx cats (the bald ones) which is both hilarious and logical since both are small, have short hair, big eyes... The tree doesn't just cluster similar things it also shows you how dissimilar things are connected.

Because the layout is a tree, you get operations that point clouds can't support:

path = model.path(idx_a, idx_b)        # nodes along the tree path
d = model.distance(idx_a, idx_b)        # sum of edge weights along the path
pseudotime = model.distances_from(idx)  # tree distance from one point to all others

Installation

pip install tmap2

Optional extras:

pip install rdkit # chemistry helpers (fingerprints_from_smiles, molecular_properties)
pip install jupyter-scatter # notebook interactive widgets
pip install biopython # protein helpers (ProtParam properties, PDB parsing)

Note: The import name is tmap, not tmap2.

Quick Start

Binary Data (e.g. Chemical Fingerprints)

from tmap.utils import fingerprints_from_smiles
from tmap import TMAP

smiles = [...] # Your smiles list
# Get Binary fingerprints (Need Jaccard distance)
fps = fingerprints_from_smiles(smiles, fp_type="morgan", radius=2, n_bits=2048)
model = TMAP(metric="jaccard", n_neighbors=20, seed=42).fit(fps)
model.write_html("map.html") # Save in html file
# model.show() # See in Jupyter Notebook 

Continuous Vectors (e.g. Protein Embeddings)

# embeddings (use cosine / euclidean distances)
X = np.random.random((1000, 128)).astype(np.float32)
model = TMAP(metric="cosine", n_neighbors=20).fit(X)
# model.write_html("tmap.html") # Save in html file
model.show() # See in Jupyter Notebook 

Key Features

• Tree structure: follow branches, trace paths, compute pseudotime
• Deterministic: same input + seed = same output
• Multiple metrics: jaccard, cosine, euclidean, precomputed
• Incremental: add_points() and transform() for adding new data into an existing TMAP
• Model persistence: save() / load()
• Three viz backends: interactive HTML, jupyter-scatter, matplotlib

Visualization (add colors, labels...)

Notebook widgets: color switching, categorical filtering, and lasso selection with pandas-backed metadata:

Add Colors & Labels

Adding colors is quite simple. Just pass the name of the layout (e.g. Molecular Weight, Age, Protein Lenght ...), a list of values for each node and matplotlib color. If the data is categorical (e.g. Age or Heavy Atom Count) pass categorical=True so that categorical colors like tab10 become available. To add labels (i.e. data that is not needed for coloring the nodes) just pass a name for the labels and the list of values.

model = TMAP(metric="jaccard").fit(X)
viz = model.to_tmapviz() 
viz.add_color_layout("Molecular Weight", mw.tolist(), categorical=False) 
viz.add_color_layout("Scaffold", scaffolds, categorical=True, color="tab10")
viz.add_label("SMILES", smiles_list)
viz.show(width=1000, height=620, controls=True) # to see in jupyter notebook
# viz.write_html("mytmap.html") # to save and see as HTML in the browser

Here SMILES are added as label which will not trigger the 2D image of the structure. If you want to see the structures add smiles via add_smiles(smiles_list)

If you save using viz.write_html("name.html") the Interactive HTML becomes available which supports lasso selection, light/dark theme, filter and search panels, pinned metadata cards, binary mode for large datasets.

Alternatively, you can see it with matplotlib by using Static plots matplotlib for publication figures: model.plot_static(color_by=labels)

Domain Utilities

Built-in helpers for common scientific workflows:

from tmap.utils.chemistry import fingerprints_from_smiles, molecular_properties
from tmap.utils.proteins import fetch_uniprot, sequence_properties
from tmap.utils.singlecell import from_anndata

Notebooks

Lower-Level Pipeline

For direct control over indexing, hashing, and layout, see the legacy pipeline notebook. The main building blocks:

from tmap.index import USearchIndex           # dense / binary kNN
from tmap import MinHash, LSHForest           # Jaccard on sets / strings
from tmap.layout import LayoutConfig, layout_from_lsh_forest

Your Data
   ├─→ Binary matrix ─────────→ USearch        (Jaccard / cosine / euclidean)
   └─→ Sets / strings ───────→ MinHash → LSHForest
                ↓
             k-NN Graph → MST → OGDF Tree Layout → Interactive Visualization

Development

git clone https://github.com/afloresep/tmap2.git
cd tmap2
pip install ".[dev]"
pytest -v

License

MIT License; see LICENSE for details.

Citation

If you use TMAP2 in your research, please cite:

From Proteins and Molecules to Cats and Dogs: Visualization as Scalable Minimum Spanning Trees
Alejandro Flores Sepúlveda, Maarten Boneschansker, Daniel Probst, Jean-Louis Reymond
ChemRxiv, 2026.
https://doi.org/10.26434/chemrxiv.15008307/v1

@article{floressepulveda2026tmap,
  title   = {From Proteins and Molecules to Cats and Dogs: Visualization as Scalable Minimum Spanning Trees},
  author  = {Flores Sepúlveda, Alejandro and Boneschansker, Maarten and Probst, Daniel and Reymond, Jean-Louis},
  year    = {2026},
  doi     = {10.26434/chemrxiv.15008307/v1},
  url     = {https://doi.org/10.26434/chemrxiv.15008307/v1},
  note    = {ChemRxiv preprint}
}