Datasets:

lagosproject
/

ALPHAGenome-Embeddings

chrom stringclasses 24 values	start int64 0 249M	end int64 131k 249M
chr10	0	131,072
chr10	131,072	262,144
chr10	262,144	393,216
chr10	393,216	524,288
chr10	524,288	655,360
chr10	655,360	786,432
chr10	786,432	917,504
chr10	917,504	1,048,576
chr10	1,048,576	1,179,648
chr10	1,179,648	1,310,720
chr10	1,310,720	1,441,792
chr10	1,441,792	1,572,864
chr10	1,572,864	1,703,936
chr10	1,703,936	1,835,008
chr10	1,835,008	1,966,080
chr10	1,966,080	2,097,152
chr10	2,097,152	2,228,224
chr10	2,228,224	2,359,296
chr10	2,359,296	2,490,368
chr10	2,490,368	2,621,440
chr10	2,621,440	2,752,512
chr10	2,752,512	2,883,584
chr10	2,883,584	3,014,656
chr10	3,014,656	3,145,728
chr10	3,145,728	3,276,800
chr10	3,276,800	3,407,872
chr10	3,407,872	3,538,944
chr10	3,538,944	3,670,016
chr10	3,670,016	3,801,088
chr10	3,801,088	3,932,160
chr10	3,932,160	4,063,232
chr10	4,063,232	4,194,304
chr10	4,194,304	4,325,376
chr10	4,325,376	4,456,448
chr10	4,456,448	4,587,520
chr10	4,587,520	4,718,592
chr10	4,718,592	4,849,664
chr10	4,849,664	4,980,736
chr10	4,980,736	5,111,808
chr10	5,111,808	5,242,880
chr10	5,242,880	5,373,952
chr10	5,373,952	5,505,024
chr10	5,505,024	5,636,096
chr10	5,636,096	5,767,168
chr10	5,767,168	5,898,240
chr10	5,898,240	6,029,312
chr10	6,029,312	6,160,384
chr10	6,160,384	6,291,456
chr10	6,291,456	6,422,528
chr10	6,422,528	6,553,600
chr10	6,553,600	6,684,672
chr10	6,684,672	6,815,744
chr10	6,815,744	6,946,816
chr10	6,946,816	7,077,888
chr10	7,077,888	7,208,960
chr10	7,208,960	7,340,032
chr10	7,340,032	7,471,104
chr10	7,471,104	7,602,176
chr10	7,602,176	7,733,248
chr10	7,733,248	7,864,320
chr10	7,864,320	7,995,392
chr10	7,995,392	8,126,464
chr10	8,126,464	8,257,536
chr10	8,257,536	8,388,608
chr10	8,388,608	8,519,680
chr10	8,519,680	8,650,752
chr10	8,650,752	8,781,824
chr10	8,781,824	8,912,896
chr10	8,912,896	9,043,968
chr10	9,043,968	9,175,040
chr10	9,175,040	9,306,112
chr10	9,306,112	9,437,184
chr10	9,437,184	9,568,256
chr10	9,568,256	9,699,328
chr10	9,699,328	9,830,400
chr10	9,830,400	9,961,472
chr10	9,961,472	10,092,544
chr10	10,092,544	10,223,616
chr10	10,223,616	10,354,688
chr10	10,354,688	10,485,760
chr10	10,485,760	10,616,832
chr10	10,616,832	10,747,904
chr10	10,747,904	10,878,976
chr10	10,878,976	11,010,048
chr10	11,010,048	11,141,120
chr10	11,141,120	11,272,192
chr10	11,272,192	11,403,264
chr10	11,403,264	11,534,336
chr10	11,534,336	11,665,408
chr10	11,665,408	11,796,480
chr10	11,796,480	11,927,552
chr10	11,927,552	12,058,624
chr10	12,058,624	12,189,696
chr10	12,189,696	12,320,768
chr10	12,320,768	12,451,840
chr10	12,451,840	12,582,912
chr10	12,582,912	12,713,984
chr10	12,713,984	12,845,056
chr10	12,845,056	12,976,128
chr10	12,976,128	13,107,200

End of preview. Expand in Data Studio

ALPHAGenome hg38 Embeddings

Pre-computed ALPHAGenome DNA foundation model embeddings for the entire human genome (hg38 / GRCh38).

The human genome is divided into ~22,000 non-overlapping 131 KB bins. Each bin's DNA sequence is embedded into a 3,072-dimensional latent space using the ALPHAGenome foundation model.

Companion project

These embeddings power the ALPHAGenome UMAP Explorer — an interactive browser visualization of latent relationships between genomic regions:

→ GitHub: DNAEmbeddings
→ Live Demo

Scripts to download this dataset and re-run the UMAP pipeline are included in that repo.

Dataset structure

data/
  chr1_embeddings.npy     # float32 array, shape (N, 3072)
  chr1_metadata.csv       # columns: chrom, start, end
  chr2_embeddings.npy
  chr2_metadata.csv
  ...                     # one pair per chromosome (chr1–22, chrX, chrY)

Each .npy file and its paired .csv share the same row order — row i in the embeddings corresponds to row i in the metadata.

Quick usage

from huggingface_hub import hf_hub_download
import numpy as np
import pandas as pd

# Download a single chromosome
emb_path  = hf_hub_download("lagosproject/ALPHAGenome-Embeddings", "data/chr1_embeddings.npy", repo_type="dataset")
meta_path = hf_hub_download("lagosproject/ALPHAGenome-Embeddings", "data/chr1_metadata.csv",    repo_type="dataset")

embeddings = np.load(emb_path)          # (1750, 3072)
metadata   = pd.read_csv(meta_path)     # chrom | start | end

Or download everything at once (≈312 MB):

from huggingface_hub import snapshot_download
snapshot_download("lagosproject/ALPHAGenome-Embeddings", repo_type="dataset", local_dir="res/")

Statistics

Chromosome	Windows	Embedding shape
chr1	1,750	(1750, 3072)
chr2	1,829	(1829, 3072)
chr3	1,509	(1509, 3072)
chr4	1,439	(1439, 3072)
chr5	1,379	(1379, 3072)
chr6	1,294	(1294, 3072)
chr7	1,206	(1206, 3072)
chr8	1,100	(1100, 3072)
chr9	915	(915, 3072)
chr10	1,013	(1013, 3072)
chr11	1,022	(1022, 3072)
chr12	1,012	(1012, 3072)
chr13	743	(743, 3072)
chr14	687	(687, 3072)
chr15	640	(640, 3072)
chr16	619	(619, 3072)
chr17	627	(627, 3072)
chr18	609	(609, 3072)
chr19	444	(444, 3072)
chr20	482	(482, 3072)
chr21	288	(288, 3072)
chr22	285	(285, 3072)
chrX	1,170	(1170, 3072)
chrY	191	(191, 3072)
Total	22,253	3,072-dim

Data sources

Source	Usage
ALPHAGenome	DNA sequence embeddings (3,072-dim per 131 KB bin)
UCSC hg38	Reference genome coordinates

License

CC BY 4.0 — free to use for research and educational purposes with attribution.

Downloads last month: 111