Construct lienage tree using Trisicell-Boost

This example shows how to construct a lineage tree using Trisicell-Boost on a binary single-cell genotype matrix.

import trisicell as tsc

# sphinx_gallery_thumbnail_path = "_static/thumbnails/trisicell-boost.png"

First, we load a binary test single-cell genotype data.

df_in = tsc.datasets.test()
df_in.head()

	mut0	mut1	mut2	mut3	mut4	mut5	mut6	mut7	mut8	mut9	mut10	mut11	mut12	mut13	mut14	mut15	mut16	mut17	mut18	mut19
cellIDxmutID
cell0	1	1	1	1	1	1	1	0	3	3	0	3	0	0	0	3	0	0	0	0
cell1	3	0	0	0	0	0	3	0	1	1	3	1	0	0	3	0	0	1	1	1
cell2	0	0	0	0	0	0	0	1	1	3	1	1	0	0	3	1	1	1	1	0
cell3	0	0	0	0	0	0	0	3	1	0	0	1	0	0	0	1	1	1	0	0
cell4	0	0	0	0	0	0	0	1	1	1	1	1	0	0	3	0	0	0	0	0

Next, using trisicell.tl.booster() we remove the single-cell noises from the input.

df_out = tsc.tl.booster(
    df_in,
    alpha=0.0000001,
    beta=0.1,
    solver="PhISCS",
    sample_on="muts",
    sample_size=15,
    n_samples=88,
    begin_index=0,
    n_jobs=1,
    dep_weight=50,
)
df_out.head()

SUBSAMPLING    (1/3):   0%|                                                  | 0/88 [00:00<?, ?it/s]
SUBSAMPLING    (1/3):   0%|                                                  | 0/88 [00:03<?, ?it/s]

DEPENDENCIES   (2/3):   0%|                                                  | 0/88 [00:00<?, ?it/s]
DEPENDENCIES   (2/3):  88%|###################################     | 77/88 [00:00<00:00, 765.62it/s]
DEPENDENCIES   (2/3): 100%|########################################| 88/88 [00:00<00:00, 765.13it/s]

RECONSTRUCTION (3/3):   0%|                                                  | 0/20 [00:00<?, ?it/s]
RECONSTRUCTION (3/3): 100%|######################################| 20/20 [00:00<00:00, 11781.75it/s]
input -- size: 20x20
input -- 0: 226#, 56.5%
input -- 1: 94#, 23.5%
input -- NA: 80#, 20.0%
input -- CF: False
output -- size: 20x20
output -- 0: 278#, 69.5%
output -- 1: 122#, 30.5%
output -- NA: 0#, 0.0%
output -- CF: True
output -- time: 3.4s (0:00:03.419101)
flips -- #0->1: 10
flips -- #1->0: 0
flips -- #NA->0: 62
flips -- #NA->1: 18
rates -- FN: 0.096
rates -- FP: 0.00000000
rates -- NA: 0.200
score -- NLL: 32.92976100177747
TREE_SCORE: 179.1110811296527

	mut0	mut1	mut2	mut3	mut4	mut5	mut6	mut7	mut8	mut9	mut10	mut11	mut12	mut13	mut14	mut15	mut16	mut17	mut18	mut19
cellIDxmutID
cell0	1	1	1	1	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0
cell1	0	0	0	0	0	0	0	1	1	1	1	1	0	0	0	1	1	1	1	1
cell2	0	0	0	0	0	0	0	1	1	1	1	1	0	0	0	1	1	1	1	0
cell3	0	0	0	0	0	0	0	0	1	1	0	1	0	0	0	1	1	1	0	0
cell4	0	0	0	0	0	0	0	1	1	1	1	1	0	0	0	1	1	1	0	0

Finally, using trisicell.ul.is_conflict_free_gusfield() we check whether the inferred genotype matrix is conflict-free or not.

is_cf = tsc.ul.is_conflict_free_gusfield(df_out)
is_cf

True

Estimated memory usage: 15 MB