Dataset analysis: `dms`

Run id 20260515_data_dms_seed0 · 2026-05-18T11:40:10+00:00 · championsnet

TL;DR

324,092 rows / 324,092 unique sequences (0 duplicates).
Mean labels per row: 2.36 / 8. Rows with all 8 labels: 2,240. Rows with zero labels: 1,546.
WT length rate: 100.00%; canonical AA rate: 100.00%; N-glyco motif rate: 3.17%.
Mean edit distance to WT: 2.33 (median 2, p99 4).
Overall per-position entropy: 0.11 bits; 0/232 positions above 1 bit.

Basic

rows	324,092
columns	80
unique_seq_hash	324,092
duplicate_rows	0

source_dataset distribution

source	rows	fraction
`dms`	324,092	100.00%

label_source distribution

label_source	rows	fraction
`sort_label`	324,092	100.00%

Label coverage

Labels per row

labels per row	rows	fraction	distribution
0	1,546	0.48%	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
1	120,143	37.07%	████████████████████████████████
2	83,685	25.82%	██████████████████████░░░░░░░░░░
3	53,627	16.55%	██████████████░░░░░░░░░░░░░░░░░░
4	32,352	9.98%	████████░░░░░░░░░░░░░░░░░░░░░░░░
5	17,729	5.47%	████░░░░░░░░░░░░░░░░░░░░░░░░░░░░
6	8,654	2.67%	██░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
7	4,116	1.27%	█░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
8	2,240	0.69%	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░

Per-receptor balance

Receptor	binders	non-binders	missing	missing %	bind % (of called)
`R1`	44,290	57,765	222,037	68.51%	43.40%
`FcRn6`	48,488	32,800	242,804	74.92%	59.65%
`R2AH`	31,224	44,186	248,682	76.73%	41.41%
`R2AR`	54,329	54,148	215,615	66.53%	50.08%
`R2B`	71,033	68,194	184,865	57.04%	51.02%
`R3AF`	23,593	32,052	268,447	82.83%	42.40%
`R3AV`	23,824	27,572	272,696	84.14%	46.35%
`R3B`	40,826	110,779	172,487	53.22%	26.93%

Profile strings

3,377 unique profile strings; 2,240 rows (0.69%) have all 8 receptors labelled.

Top 20 most common profiles

profile	rows	fraction
`-------0`	33,576	10.36%
`-0------`	15,632	4.82%
`----0---`	13,362	4.12%
`---0----`	10,448	3.22%
`0-------`	8,046	2.48%
`---1----`	7,002	2.16%
`----1---`	6,848	2.11%
`0------0`	5,771	1.78%
`----1--0`	5,158	1.59%
`--0-----`	4,735	1.46%
`-------1`	4,169	1.29%
`0---0---`	4,102	1.27%
`----0--0`	3,909	1.21%
`---00---`	3,778	1.17%
`----1--1`	3,651	1.13%
`-1-----0`	2,924	0.90%
`0---0--0`	2,706	0.83%
`--1-----`	2,636	0.81%
`-1------`	2,577	0.80%
`-----0--`	2,575	0.79%

Sequence features

wt_length_rate	1.00
canonical_aa_rate	1.00
n_glyco_motif_rate	0.0317
in_variable_region_only_rate	0.5407

Edit distance to WT

mean = 2.33, std = 1.08, median = 2, p90 = 3, p99 = 4, max = 61.

edit distance	rows	histogram
`0`	1	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`1`	3,729	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`2`	235,839	██████████████████████████████
`3-4`	81,828	██████████░░░░░░░░░░░░░░░░░░░░
`5-7`	1,767	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`8-9`	190	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`10-14`	147	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`15-19`	294	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`20-29`	270	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`30-49`	11	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`>=50`	16	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░

Mean edit distance, by per-receptor label

Receptor	binders	non-binders	missing
`R1`	mean=2.02, median=2 (n=44,290)	mean=2.36, median=2 (n=57,765)	mean=2.39, median=2 (n=222,037)
`FcRn6`	mean=2.01, median=2 (n=48,488)	mean=2.44, median=2 (n=32,800)	mean=2.38, median=2 (n=242,804)
`R2AH`	mean=2.01, median=2 (n=31,224)	mean=2.29, median=2 (n=44,186)	mean=2.38, median=2 (n=248,682)
`R2AR`	mean=2.07, median=2 (n=54,329)	mean=2.43, median=2 (n=54,148)	mean=2.38, median=2 (n=215,615)
`R2B`	mean=2.08, median=2 (n=71,033)	mean=2.42, median=2 (n=68,194)	mean=2.40, median=2 (n=184,865)
`R3AF`	mean=1.99, median=2 (n=23,593)	mean=2.29, median=2 (n=32,052)	mean=2.37, median=2 (n=268,447)
`R3AV`	mean=1.98, median=2 (n=23,824)	mean=2.31, median=2 (n=27,572)	mean=2.37, median=2 (n=272,696)
`R3B`	mean=2.09, median=2 (n=40,826)	mean=2.24, median=2 (n=110,779)	mean=2.45, median=2 (n=172,487)

Read counts per receptor

Receptor	n with count	mean	median	p90	p99	max
`R1`	104,224	38.30	2	10	335	1,409,172
`FcRn6`	82,366	47.38	2	14	387	1,245,548
`R2AH`	77,747	14.44	2	7	175	296,845
`R2AR`	111,095	19.16	2	9	229	601,641
`R2B`	142,854	28.03	2	10	270	996,987
`R3AF`	57,426	10.54	2	8	122	106,112
`R3AV`	52,773	11.08	2	8	120	135,940
`R3B`	165,755	29.67	2	9	296	1,718,201

Sequence diversity (per-position Shannon entropy)

Overall (sampled): n=200,000, mean entropy = 0.11 bits, max position = 0.58, positions above 1 bit = 0/232, zero-entropy positions = 14.

Sequences with all 8 labels: n=2,240, mean entropy = 0.07 bits, max position = 0.44, positions above 1 bit = 0/232, zero-entropy positions = 14.

Per receptor (binders vs non-binders)

Receptor	binders	non-binders
`R1`	n=44,290, mean H = 0.09	n=57,765, mean H = 0.10
`FcRn6`	n=48,488, mean H = 0.09	n=32,800, mean H = 0.10
`R2AH`	n=31,224, mean H = 0.09	n=44,186, mean H = 0.10
`R2AR`	n=54,329, mean H = 0.10	n=54,148, mean H = 0.10
`R2B`	n=71,033, mean H = 0.10	n=68,194, mean H = 0.10
`R3AF`	n=23,593, mean H = 0.09	n=32,052, mean H = 0.10
`R3AV`	n=23,824, mean H = 0.09	n=27,572, mean H = 0.10
`R3B`	n=40,826, mean H = 0.10	n=100,000, mean H = 0.10

Generated by scripts/data/analyze_dataset.py. Machine-readable companion: analysis.yaml in the same run directory.

Dataset analysis: dms