Dataset analysis: `paper_refiltered`

Run id 20260515_data_paper_refiltered_seed0 · 2026-05-18T11:42:14+00:00 · championsnet

TL;DR

1,108,275 rows / 1,108,275 unique sequences (0 duplicates).
Mean labels per row: 1.67 / 8. Rows with all 8 labels: 4,458. Rows with zero labels: 174.
WT length rate: 100.00%; canonical AA rate: 100.00%; N-glyco motif rate: 17.59%.
Mean edit distance to WT: 6.12 (median 6, p99 15).
Overall per-position entropy: 0.18 bits; 13/232 positions above 1 bit.

Basic

rows	1,108,275
columns	80
unique_seq_hash	1,108,275
duplicate_rows	0

source_dataset distribution

source	rows	fraction
`paper_refiltered`	1,108,275	100.00%

label_source distribution

label_source	rows	fraction
`sort_label`	1,108,275	100.00%

Label coverage

Labels per row

labels per row	rows	fraction	distribution
0	174	0.02%	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
1	701,541	63.30%	████████████████████████████████
2	230,657	20.81%	██████████░░░░░░░░░░░░░░░░░░░░░░
3	92,822	8.38%	████░░░░░░░░░░░░░░░░░░░░░░░░░░░░
4	42,158	3.80%	█░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
5	20,477	1.85%	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
6	10,351	0.93%	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
7	5,637	0.51%	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
8	4,458	0.40%	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░

Per-receptor balance

Receptor	binders	non-binders	missing	missing %	bind % (of called)
`R1`	124,861	173,163	810,251	73.11%	41.90%
`FcRn6`	98,560	191,502	818,213	73.83%	33.98%
`R2AH`	16,050	134,440	957,785	86.42%	10.67%
`R2AR`	22,756	185,742	899,777	81.19%	10.91%
`R2B`	16,081	138,384	953,810	86.06%	10.41%
`R3AF`	14,381	218,313	875,581	79.00%	6.18%
`R3AV`	21,198	215,112	871,965	78.68%	8.97%
`R3B`	10,801	268,223	829,251	74.82%	3.87%

Profile strings

2,287 unique profile strings; 4,458 rows (0.40%) have all 8 receptors labelled.

Top 20 most common profiles

profile	rows	fraction
`-------0`	108,554	9.79%
`0-------`	88,200	7.96%
`------0-`	87,396	7.89%
`-----0--`	84,154	7.59%
`-0------`	76,317	6.89%
`---0----`	74,428	6.72%
`--0-----`	49,137	4.43%
`----0---`	48,743	4.40%
`-1------`	36,074	3.26%
`1-------`	33,971	3.07%
`-0-----0`	10,564	0.95%
`-----0-0`	10,300	0.93%
`------00`	9,707	0.88%
`10------`	7,956	0.72%
`-----00-`	7,814	0.71%
`---0---0`	7,733	0.70%
`-0---0--`	7,579	0.68%
`0------0`	7,520	0.68%
`1------0`	6,885	0.62%
`-0----0-`	6,721	0.61%

Sequence features

wt_length_rate	1.00
canonical_aa_rate	1.00
n_glyco_motif_rate	0.1759
in_variable_region_only_rate	1.00

Edit distance to WT

mean = 6.12, std = 2.33, median = 6, p90 = 8, p99 = 15, max = 56.

edit distance	rows	histogram
`0`	1	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`1`	1,122	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`2`	23,433	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`3-4`	127,645	████░░░░░░░░░░░░░░░░░░░░░░░░░░
`5-7`	808,987	██████████████████████████████
`8-9`	73,779	██░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`10-14`	58,391	██░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`15-19`	11,072	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`20-29`	3,717	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`30-49`	124	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
`>=50`	4	░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░

Mean edit distance, by per-receptor label

Receptor	binders	non-binders	missing
`R1`	mean=5.46, median=6 (n=124,861)	mean=6.00, median=6 (n=173,163)	mean=6.24, median=6 (n=810,251)
`FcRn6`	mean=4.96, median=5 (n=98,560)	mean=6.11, median=6 (n=191,502)	mean=6.26, median=6 (n=818,213)
`R2AH`	mean=3.64, median=4 (n=16,050)	mean=5.74, median=6 (n=134,440)	mean=6.21, median=6 (n=957,785)
`R2AR`	mean=3.88, median=4 (n=22,756)	mean=5.89, median=6 (n=185,742)	mean=6.22, median=6 (n=899,777)
`R2B`	mean=3.50, median=3 (n=16,081)	mean=5.81, median=6 (n=138,384)	mean=6.20, median=6 (n=953,810)
`R3AF`	mean=3.48, median=3 (n=14,381)	mean=5.86, median=6 (n=218,313)	mean=6.22, median=6 (n=875,581)
`R3AV`	mean=3.86, median=4 (n=21,198)	mean=5.88, median=6 (n=215,112)	mean=6.23, median=6 (n=871,965)
`R3B`	mean=3.58, median=4 (n=10,801)	mean=5.82, median=6 (n=268,223)	mean=6.24, median=6 (n=829,251)

Read counts per receptor

Receptor	n with count	mean	median	p90	p99	max
`R1`	298,330	6.11	3	7	19	402,232
`FcRn6`	290,459	6.19	3	7	19	418,715
`R2AH`	150,845	10.86	2	5	31	884,406
`R2AR`	208,937	8.90	2	6	23	865,757
`R2B`	154,646	6.25	2	5	24	337,909
`R3AF`	233,074	6.37	2	6	18	433,711
`R3AV`	236,617	8.17	2	6	20	770,365
`R3B`	279,790	7.14	3	7	20	591,241

Sequence diversity (per-position Shannon entropy)

Overall (sampled): n=200,000, mean entropy = 0.18 bits, max position = 2.89, positions above 1 bit = 13/232, zero-entropy positions = 95.

Sequences with all 8 labels: n=4,458, mean entropy = 0.09 bits, max position = 2.68, positions above 1 bit = 5/232, zero-entropy positions = 95.

Per receptor (binders vs non-binders)

Receptor	binders	non-binders
`R1`	n=100,000, mean H = 0.11	n=100,000, mean H = 0.17
`FcRn6`	n=98,560, mean H = 0.12	n=100,000, mean H = 0.15
`R2AH`	n=16,050, mean H = 0.11	n=100,000, mean H = 0.17
`R2AR`	n=22,756, mean H = 0.11	n=100,000, mean H = 0.17
`R2B`	n=16,081, mean H = 0.11	n=100,000, mean H = 0.17
`R3AF`	n=14,381, mean H = 0.10	n=100,000, mean H = 0.17
`R3AV`	n=21,198, mean H = 0.11	n=100,000, mean H = 0.17
`R3B`	n=10,801, mean H = 0.10	n=100,000, mean H = 0.17

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

Dataset analysis: paper_refiltered