Benchmark transcription rate prediction

Benchmark transcription rate prediction#

Notebooks for benchmarking transcription rate prediction on metabolic labelled cell cycle datasets

Library imports#

import numpy as np
import pandas as pd
import scipy
import torch
from scipy.stats import ttest_ind

import matplotlib.pyplot as plt
import mplscience
import seaborn as sns

import scanpy as sc
from regvelo import REGVELOVI

from rgv_tools import DATA_DIR, FIG_DIR
from rgv_tools.plotting._significance import add_significance, get_significance

Constants#

DATASET = "cell_cycle_rpe1"

SAVE_DATA = True
SAVE_FIGURES = True
if SAVE_DATA:
    (DATA_DIR / DATASET / "processed").mkdir(parents=True, exist_ok=True)
    (FIG_DIR / DATASET).mkdir(parents=True, exist_ok=True)

Data loading#

adata = sc.read_h5ad(DATA_DIR / DATASET / "processed" / "adata_processed.h5ad")

alpha = pd.read_csv(
    DATA_DIR / DATASET / "raw" / "aax3072_table-s1.csv", index_col=2, header=1
)  # downloaded from the supplementary of https://www.science.org/doi/10.1126/science.aax3072

alpha = alpha.filter(regex="^norm_kappa_")
alpha
norm_kappa_000 norm_kappa_001 norm_kappa_002 norm_kappa_003 norm_kappa_004 norm_kappa_005 norm_kappa_006 norm_kappa_007 norm_kappa_008 norm_kappa_009 ... norm_kappa_291 norm_kappa_292 norm_kappa_293 norm_kappa_294 norm_kappa_295 norm_kappa_296 norm_kappa_297 norm_kappa_298 norm_kappa_299 norm_kappa_300
gene_symbol
UBALD2 1.179867 1.179867 1.179867 1.179867 1.179867 1.179867 1.165868 1.165868 1.124469 1.124469 ... 1.393822 1.389139 1.387780 1.387780 1.387780 1.387780 1.387780 1.387780 1.387780 1.387780
LMNB2 0.118109 0.118109 0.118109 0.091849 0.091849 0.091849 0.091849 0.091849 0.091849 0.016334 ... 0.328110 0.328110 0.328110 0.328110 0.328110 0.320427 0.320427 0.320427 0.304818 0.304818
SOGA1 0.088258 0.088258 0.020962 0.020962 0.020962 0.020962 0.012680 0.012680 0.012680 0.012680 ... 0.518963 0.470260 0.432552 0.432552 0.397859 0.397859 0.381700 0.381700 0.381700 0.381700
PPIP5K2 0.108402 0.071083 0.071083 0.071083 0.071083 0.071083 0.071083 0.071083 0.071083 0.043664 ... 0.593905 0.593905 0.457897 0.457897 0.457897 0.457897 0.457897 0.447648 0.447648 0.447648
PTPRJ 1.002031 1.002031 1.002031 1.002031 1.002031 1.002031 1.002031 1.002031 0.888077 0.861586 ... 1.276865 1.270956 1.266449 1.266449 1.266449 1.251248 1.251248 1.251248 1.251248 1.251248
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
SGMS2 -1.192674 -1.192674 -1.192674 -1.192674 -1.208510 -1.224163 -1.208510 -1.208510 -1.208510 -1.208510 ... -0.453817 -0.479000 -0.479000 -0.479000 -0.479000 -0.504759 -0.504759 -0.542460 -0.542460 -0.549843
PMAIP1 -0.183754 -0.183754 -0.183754 -0.183754 -0.183754 -0.183754 -0.183754 -0.183754 -0.184812 -0.223222 ... -0.042112 -0.044428 -0.044428 -0.060940 -0.060940 -0.060940 -0.080172 -0.080172 -0.080172 -0.080172
DHCR24 -0.450735 -0.450735 -0.450735 -0.450735 -0.450735 -0.450735 -0.450735 -0.450735 -0.484615 -0.493553 ... -0.320440 -0.329280 -0.329280 -0.329280 -0.329280 -0.329280 -0.330464 -0.330464 -0.376690 -0.384143
POLE -0.529833 -0.529833 -0.529833 -0.575677 -0.575677 -0.575677 -0.575677 -0.575677 -0.603306 -0.603306 ... -0.412773 -0.420573 -0.420573 -0.420573 -0.488070 -0.506770 -0.506770 -0.506770 -0.506770 -0.506770
ANKRD36C -0.432021 -0.432021 -0.432021 -0.432021 -0.432021 -0.432021 -0.447457 -0.447457 -0.467636 -0.498658 ... -0.221038 -0.256899 -0.256899 -0.274694 -0.274694 -0.274694 -0.274694 -0.274694 -0.274694 -0.274694

528 rows × 301 columns

alpha.columns = np.array([i.replace("norm_kappa_", "") for i in alpha.columns], dtype=np.float32)

pos = adata.obs["cell_cycle_position"].unique()
GEX = []

for i in pos:
    gex = adata[adata.obs["cell_cycle_position"] == i].X.A.mean(axis=0)
    GEX.append(gex)

GEX = np.array(GEX).T

GEX = pd.DataFrame(GEX, index=adata.var_names.tolist())

GEX
0 1 2 3 4 5 6 7 8 9 ... 280 281 282 283 284 285 286 287 288 289
DEPDC1 0.161310 0.328377 0.126236 0.000000 0.445840 0.742209 0.684013 0.317394 0.509439 0.180416 ... 0.337324 0.523796 1.649758 0.000000 0.357087 1.076477 0.000000 0.000000 0.000000 0.000000
TRIO 1.981804 1.515634 1.730185 2.096108 1.660738 1.462572 1.611647 2.059388 2.121015 1.883820 ... 1.791328 2.165788 0.876263 2.176786 0.636869 2.048151 2.324368 2.076471 2.328215 1.574311
ADAMTS6 1.273340 1.115689 0.793328 0.815586 0.724796 0.495016 0.865061 1.060191 1.100434 0.940455 ... 1.411591 1.539369 0.876263 1.219352 0.525509 1.306756 1.440009 1.383749 0.000000 0.000000
POLQ 0.220024 0.093262 0.655998 0.120990 0.369249 0.329390 0.096619 0.529976 0.502093 0.098990 ... 0.309368 0.322715 1.888225 0.260449 0.000000 0.230847 0.470612 0.000000 0.320231 1.574311
SYNE2 0.456051 0.121919 0.452270 0.317492 0.821573 0.516719 0.412208 0.884521 0.536059 0.589152 ... 0.966696 0.376468 1.336004 1.522307 0.000000 0.851015 0.000000 0.691451 0.320231 1.069426
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
MANCR 1.107140 0.836700 0.756655 0.554660 0.632607 1.029576 0.840048 1.151329 1.099359 0.443124 ... 0.780349 1.278557 0.000000 1.791222 0.335229 0.574196 1.046393 1.096350 0.320231 0.000000
MTATP6P1 2.288898 2.405079 2.159031 2.227160 2.321663 2.195483 2.659615 2.515312 2.502668 2.483743 ... 2.828326 2.249292 2.709321 2.271032 2.617242 2.023398 1.604221 2.830018 2.296170 2.158117
MIR100HG 1.508564 1.982803 1.997843 2.543906 2.180073 2.280532 2.287652 1.908947 1.789601 2.454187 ... 1.833768 2.028000 0.876263 1.520726 2.845396 2.284718 1.737294 1.943000 2.607786 1.908226
SNHG1 1.340924 0.865615 0.407378 0.911950 1.100356 0.833532 1.037473 0.792740 1.336403 0.926061 ... 0.710278 0.952297 1.336004 0.341737 0.692316 0.947387 1.179466 1.096350 0.653094 1.908226
SNHG19 0.349352 0.636762 0.260200 0.517246 0.360954 0.710418 0.619507 0.398636 0.000000 0.374218 ... 0.149108 0.376468 0.000000 0.685875 0.357087 0.919150 0.575781 0.000000 0.792144 0.000000

141 rows × 290 columns

gs = list(set(alpha.index.tolist()).intersection(adata.var_names.tolist()))

GEX.loc[gs, :]
0 1 2 3 4 5 6 7 8 9 ... 280 281 282 283 284 285 286 287 288 289
PAPPA 0.236026 0.595305 0.754837 0.815792 0.590248 0.554654 0.354364 0.417110 0.374995 0.716544 ... 0.861246 0.409270 0.876263 0.000000 0.370403 0.388174 0.000000 0.000000 0.792144 0.000000
RIMS2 0.000000 0.330232 0.247463 0.379518 0.510151 0.191398 0.000000 0.603839 0.530653 0.054663 ... 0.559815 0.573492 0.000000 0.430828 0.370403 0.388174 0.000000 0.691451 0.513865 0.000000
MIR100HG 1.508564 1.982803 1.997843 2.543906 2.180073 2.280532 2.287652 1.908947 1.789601 2.454187 ... 1.833768 2.028000 0.876263 1.520726 2.845396 2.284718 1.737294 1.943000 2.607786 1.908226
G2E3 0.236026 0.113038 0.136443 0.140211 0.337680 0.120133 0.133690 0.392664 0.870752 0.156803 ... 0.428223 0.381109 2.080633 0.742809 0.357087 0.000000 0.470612 0.000000 0.513865 1.574311
AURKA 0.851758 0.219980 0.172232 0.104639 0.240032 0.335769 0.474674 0.418812 0.273523 0.318951 ... 0.646679 0.824949 0.876263 0.000000 0.895912 1.030397 0.864525 0.000000 0.000000 1.069426
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
CENPF 1.120599 0.872054 0.628304 0.371293 0.910407 1.810582 1.590818 1.217998 1.629507 1.057226 ... 0.805158 1.354087 1.649758 1.357588 1.873811 1.849402 1.826995 1.788931 0.471913 2.524307
MCM10 0.129327 0.107488 0.685333 0.266811 0.399895 0.390223 0.047354 0.330642 0.387730 0.176696 ... 0.342713 0.135464 0.000000 0.000000 0.000000 0.230847 0.000000 0.000000 1.524149 0.000000
KPNA2 0.710659 0.691963 0.619302 0.493050 0.431500 0.712793 0.647899 0.648701 0.669845 0.639119 ... 0.473450 0.930215 2.080633 0.803296 0.370403 1.600532 1.545007 0.691451 0.000000 1.069426
PTTG1 1.176973 1.258670 1.011124 0.911626 0.870235 1.471961 1.531632 0.854305 1.377528 1.618404 ... 0.479520 1.270856 2.241926 1.539767 0.860738 1.503640 1.046393 1.383749 0.513865 1.069426
PKMYT1 0.940430 0.351545 0.664576 0.297624 0.800065 0.745133 0.171471 0.848509 0.276872 0.296031 ... 1.045587 0.669761 0.000000 0.430828 0.000000 0.851015 0.288744 0.691451 0.000000 1.069426

75 rows × 290 columns

GEX.columns = pos

GEX
230.0 66.0 109.0 85.0 177.0 49.0 55.0 196.0 254.0 60.0 ... 242.0 259.0 21.0 247.0 42.0 287.0 296.0 232.0 106.0 4.0
DEPDC1 0.161310 0.328377 0.126236 0.000000 0.445840 0.742209 0.684013 0.317394 0.509439 0.180416 ... 0.337324 0.523796 1.649758 0.000000 0.357087 1.076477 0.000000 0.000000 0.000000 0.000000
TRIO 1.981804 1.515634 1.730185 2.096108 1.660738 1.462572 1.611647 2.059388 2.121015 1.883820 ... 1.791328 2.165788 0.876263 2.176786 0.636869 2.048151 2.324368 2.076471 2.328215 1.574311
ADAMTS6 1.273340 1.115689 0.793328 0.815586 0.724796 0.495016 0.865061 1.060191 1.100434 0.940455 ... 1.411591 1.539369 0.876263 1.219352 0.525509 1.306756 1.440009 1.383749 0.000000 0.000000
POLQ 0.220024 0.093262 0.655998 0.120990 0.369249 0.329390 0.096619 0.529976 0.502093 0.098990 ... 0.309368 0.322715 1.888225 0.260449 0.000000 0.230847 0.470612 0.000000 0.320231 1.574311
SYNE2 0.456051 0.121919 0.452270 0.317492 0.821573 0.516719 0.412208 0.884521 0.536059 0.589152 ... 0.966696 0.376468 1.336004 1.522307 0.000000 0.851015 0.000000 0.691451 0.320231 1.069426
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
MANCR 1.107140 0.836700 0.756655 0.554660 0.632607 1.029576 0.840048 1.151329 1.099359 0.443124 ... 0.780349 1.278557 0.000000 1.791222 0.335229 0.574196 1.046393 1.096350 0.320231 0.000000
MTATP6P1 2.288898 2.405079 2.159031 2.227160 2.321663 2.195483 2.659615 2.515312 2.502668 2.483743 ... 2.828326 2.249292 2.709321 2.271032 2.617242 2.023398 1.604221 2.830018 2.296170 2.158117
MIR100HG 1.508564 1.982803 1.997843 2.543906 2.180073 2.280532 2.287652 1.908947 1.789601 2.454187 ... 1.833768 2.028000 0.876263 1.520726 2.845396 2.284718 1.737294 1.943000 2.607786 1.908226
SNHG1 1.340924 0.865615 0.407378 0.911950 1.100356 0.833532 1.037473 0.792740 1.336403 0.926061 ... 0.710278 0.952297 1.336004 0.341737 0.692316 0.947387 1.179466 1.096350 0.653094 1.908226
SNHG19 0.349352 0.636762 0.260200 0.517246 0.360954 0.710418 0.619507 0.398636 0.000000 0.374218 ... 0.149108 0.376468 0.000000 0.685875 0.357087 0.919150 0.575781 0.000000 0.792144 0.000000

141 rows × 290 columns

alpha = alpha.loc[:, pos]

alpha
230.0 66.0 109.0 85.0 177.0 49.0 55.0 196.0 254.0 60.0 ... 242.0 259.0 21.0 247.0 42.0 287.0 296.0 232.0 106.0 4.0
gene_symbol
UBALD2 0.692507 -0.774376 -1.027370 -0.854328 0.061159 -0.238411 -0.262981 0.161117 1.447728 -0.411593 ... 0.902076 1.580665 -0.175543 1.149451 -0.238411 1.589349 1.387780 0.705599 -0.953666 1.179867
LMNB2 0.529516 -0.552728 -1.304541 -1.376764 0.284558 -0.477744 -0.411589 0.402770 0.795242 -0.411589 ... 0.722385 0.761394 -0.910402 0.760184 -0.900394 0.394706 0.320427 0.536469 -1.304541 0.091849
SOGA1 0.108462 -0.680941 -0.018291 -0.328111 0.034385 -0.194473 -0.196554 -0.076837 0.652980 -0.273828 ... 0.320829 0.780893 -0.226716 0.534645 -0.249730 0.577887 0.397859 0.144626 0.000000 0.020962
PPIP5K2 0.831422 -1.025339 -0.659772 -0.762519 0.024266 -1.683484 -1.604536 0.097151 1.222013 -1.472045 ... 1.160917 1.231502 -1.231383 1.182067 -1.757942 0.912401 0.457897 0.983093 -0.672506 0.071083
PTPRJ 1.042036 -0.578932 -0.140358 -0.236434 -0.049393 -0.517368 -0.495903 0.366185 1.686175 -0.508249 ... 1.607530 1.776387 -0.549304 1.656482 -0.640710 1.329547 1.251248 1.325945 -0.131871 1.002031
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
SGMS2 1.387533 -0.082324 -0.075702 -0.145237 0.330238 -0.353837 -0.185794 0.529986 0.421661 -0.185794 ... 1.466910 -0.104414 -1.087653 1.387533 -0.507248 -0.432786 -0.504759 1.425201 -0.066528 -1.208510
PMAIP1 0.125522 -0.974546 -0.849682 -1.009706 0.368004 -1.032658 -0.979155 0.689779 0.212730 -0.966090 ... 0.136583 0.226378 -0.561242 0.163084 -1.016736 -0.001435 -0.060940 0.125522 -0.864333 -0.183754
DHCR24 0.796890 -0.606781 -0.235416 -0.580445 0.995562 -0.865113 -0.847082 0.949131 0.221086 -0.783282 ... 0.418463 0.180183 -0.789217 0.317341 -0.927062 -0.320440 -0.329280 0.704518 -0.346948 -0.450735
POLE 0.400946 -1.080653 0.264789 -0.195188 0.367359 -1.334873 -1.269773 0.367359 0.096167 -1.243465 ... 0.129791 -0.027705 -1.112352 0.129791 -1.365260 -0.336136 -0.506770 0.324208 0.232217 -0.575677
ANKRD36C 0.599263 -1.589231 -0.638207 -0.795766 0.650575 -1.684860 -1.676727 0.651835 0.366817 -1.664038 ... 0.505878 0.206309 -1.086746 0.468641 -1.616084 -0.148026 -0.274694 0.583993 -0.664289 -0.432021

528 rows × 290 columns

cor_gex = []
for i in range(len(gs)):
    cor_gex.append(scipy.stats.spearmanr(alpha.loc[gs, :].iloc[i, :], GEX.loc[gs, :].iloc[i, :])[0])

Estimate transcription rate#

vae = REGVELOVI.load(DATA_DIR / DATASET / "processed" / "regvelo_model", adata)

/home/icb/weixu.wang/miniconda3/envs/regvelo_test/lib/python3.10/site-packages/lightning/fabric/plugins/environments/slurm.py:204: The `srun` command is available on your system but is not used. HINT: If your intention is to run Lightning on SLURM, prepend your python command with `srun` like so: srun python /home/icb/weixu.wang/miniconda3/envs/regvelo_test/li ...

INFO     File /lustre/groups/ml01/workspace/weixu.wang/regvelo_revision/cell_cycle_REF/data/regvelo_model/model.pt 
         already downloaded                                                                                        

fit_s, fit_u = vae.rgv_expression_fit(n_samples=30, return_numpy=False)

s = torch.tensor(np.array(fit_s)).to("cuda:0")
alpha_pre = vae.module.v_encoder.transcription_rate(s)

alpha_pre = pd.DataFrame(alpha_pre.cpu().detach().numpy(), index=fit_s.index, columns=fit_s.columns)

alpha_pre = alpha_pre.loc[:, gs].T

alpha_pre_m = []
for i in pos:
    pre = alpha_pre.loc[:, adata.obs["cell_cycle_position"] == i].mean(axis=1)
    alpha_pre_m.append(pre)

alpha_pre_m = np.array(alpha_pre_m).T
alpha_pre_m = pd.DataFrame(alpha_pre_m, index=gs)
alpha_pre_m.columns = pos

cor_rgv = []
for i in range(75):
    cor_rgv.append(scipy.stats.spearmanr(alpha.loc[gs, :].iloc[i, :], alpha_pre_m.iloc[i, :])[0])

alpha_pre_m_rgv = alpha_pre_m.copy()

Estimating transcription rate with celldancer#

alpha_pre = pd.read_csv(DATA_DIR / DATASET / "processed" / "celldancer_alpha_estimate.csv", index_col=0)

alpha_pre = alpha_pre.loc[:, gs].T

alpha_pre_m = []
for i in pos:
    pre = alpha_pre.loc[:, adata.obs["cell_cycle_position"] == i].mean(axis=1)
    alpha_pre_m.append(pre)

alpha_pre_m = np.array(alpha_pre_m).T
alpha_pre_m = pd.DataFrame(alpha_pre_m, index=gs)
alpha_pre_m.columns = pos

alpha_pre_m
230.0 66.0 109.0 85.0 177.0 49.0 55.0 196.0 254.0 60.0 ... 242.0 259.0 21.0 247.0 42.0 287.0 296.0 232.0 106.0 4.0
PAPPA 0.067470 0.108662 0.085077 0.106062 0.065569 0.153572 0.140023 0.080240 0.067295 0.088635 ... 0.078329 0.075798 0.046987 0.096394 0.149704 0.069833 0.085413 0.065956 0.091174 0.061060
RIMS2 0.194880 0.184190 0.233063 0.188801 0.211795 0.167446 0.182517 0.176321 0.172191 0.183245 ... 0.166389 0.181412 0.190325 0.169382 0.175393 0.195688 0.174280 0.210455 0.266357 0.240317
MIR100HG 0.207153 0.234966 0.205020 0.240787 0.212764 0.291031 0.277944 0.194312 0.211806 0.244101 ... 0.193928 0.209133 0.198018 0.202107 0.297138 0.202776 0.198001 0.217856 0.183420 0.212776
G2E3 0.135669 0.104929 0.115884 0.104281 0.105922 0.073541 0.083583 0.109529 0.156124 0.101179 ... 0.112561 0.170952 0.068824 0.170462 0.065195 0.137725 0.102899 0.162377 0.138815 0.109731
AURKA 0.032485 0.015556 0.031590 0.019039 0.023794 0.006240 0.007497 0.036565 0.057842 0.010349 ... 0.043843 0.071995 0.003554 0.018232 0.005987 0.073007 0.130384 0.023054 0.015853 0.015490
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
CENPF 0.071015 0.050892 0.065156 0.063965 0.071007 0.040065 0.042347 0.065967 0.074917 0.051343 ... 0.072823 0.074915 0.049658 0.069647 0.039659 0.073996 0.072258 0.089910 0.079792 0.039903
MCM10 0.043488 0.078948 0.077092 0.073664 0.056498 0.077495 0.060162 0.047970 0.055300 0.074251 ... 0.083806 0.038094 0.129971 0.029473 0.089182 0.063064 0.076066 0.076595 0.077299 0.121573
KPNA2 0.075527 0.034534 0.047214 0.040717 0.046310 0.030415 0.031468 0.039342 0.080178 0.032182 ... 0.042713 0.096810 0.083730 0.074292 0.026279 0.060275 0.050828 0.034897 0.046984 0.012323
PTTG1 0.143616 0.150156 0.148792 0.139637 0.137904 0.147893 0.142231 0.145094 0.113857 0.120295 ... 0.155262 0.107316 0.101735 0.086086 0.111887 0.144526 0.115453 0.118170 0.128351 0.147436
PKMYT1 0.076249 0.060124 0.086998 0.051522 0.093748 0.049488 0.047053 0.088945 0.052630 0.039598 ... 0.127635 0.057120 0.032725 0.045265 0.037772 0.058293 0.043852 0.115498 0.055205 0.112760

75 rows × 290 columns

cor_cd = []
for i in range(75):
    cor_cd.append(scipy.stats.spearmanr(alpha.loc[gs, :].iloc[i, :], alpha_pre_m.iloc[i, :])[0])

np.mean(cor_cd)

0.4194778873230621

Violinplot#

dfs = []

g_df = pd.DataFrame({"Spearman correlation": cor_cd})
g_df["Method"] = "cellDancer"
dfs.append(g_df)

g_df = pd.DataFrame({"Spearman correlation": cor_rgv})
g_df["Method"] = "RegVelo"
dfs.append(g_df)

g_df = pd.DataFrame({"Spearman correlation": cor_gex})
g_df["Method"] = "GEX"
dfs.append(g_df)

df = pd.concat(dfs, axis=0)
df["Method"] = df["Method"].astype("category")

with mplscience.style_context():
    sns.set_style(style="whitegrid")
    fig, ax = plt.subplots(figsize=(3, 3))
    # pal = {"RegVelo": "#0173b2", "veloVI": "#de8f05"}

    sns.violinplot(
        data=df,
        ax=ax,
        # orient="h",
        x="Method",
        y="Spearman correlation",
        order=["RegVelo", "GEX", "cellDancer"],
        color="lightpink",
    )

    ttest_res = ttest_ind(
        cor_rgv,
        cor_gex,
        equal_var=False,
        alternative="greater",
    )
    significance = get_significance(pvalue=ttest_res.pvalue)
    add_significance(
        ax=ax,
        left=0,
        right=1,
        significance=significance,
        lw=1,
        bracket_level=1.05,
        c="k",
        level=0,
    )

    ttest_res = ttest_ind(
        cor_rgv,
        cor_cd,
        equal_var=False,
        alternative="greater",
    )
    significance = get_significance(pvalue=ttest_res.pvalue)
    add_significance(
        ax=ax,
        left=0,
        right=2,
        significance=significance,
        lw=1,
        bracket_level=1.05,
        c="k",
        level=0,
    )

    plt.xlabel("")

    if SAVE_FIGURES:
        fig.savefig(FIG_DIR / DATASET / "corr_sp.svg", format="svg", transparent=True, bbox_inches="tight")
    plt.show()
../_images/41f8fa31a250649be54e2e538814fe7376ab4ddc62d4b59602b0909ec25621b2.png
Contents