Datasets:

junidude14
/

PBISC-data

paper-releases/pbisc-diffusion/updates/20261012T130738Z-strictpairs-implementation/diffusion/dataset/server_1/model8/plans/paper_figures_strictpairs_20261012/archive/20261012T130738Z-strictpairs-implementation/snapshot/manuscript/tables/results_conditioning.tex ADDED Viewed

	@@ -0,0 +1,21 @@

+\begin{table}[t]
+\centering
+\footnotesize
+\caption{The bulk condition sets the generated disease effect. One model (base arm, part 16), one sampler seed and 50 deterministic steps; only the bulk that conditions each donor is replaced, so cell $i$ starts from the same noise in every row. Cell $r$ is the median Pearson correlation between matched cells of that row and the true-condition row. ``Closest donor'' counts, of 42 donors, whether the generated pseudobulk is nearest to the donor itself or to the donor whose bulk it was given. Two context rows give the scale of a change: independent noise and a fourfold change in sampler steps.}
+\label{tab:conditioning}
+\begin{tabular}{lrrrrrr}
+\toprule
+ & & & & & \multicolumn{2}{c}{Closest donor} \\
+\cmidrule(lr){6-7}
+Condition & $r$ & DEG & comp.\ $L_1$ & cell $r$ & self & source \\
+\midrule
+Donor's own bulk                       & $+0.241$ & 57 & 0.227 & --- & --- & --- \\
+Bulk of the opposite disease group     & $-0.272$ & 16 & 0.430 & 0.972 & 0 & 40 \\
+Bulk of another donor (derangement)    & $+0.140$ & 0  & 0.502 & 0.969 & 0 & 42 \\
+Cohort-mean bulk                       & $-0.041$ & 10 & 0.350 & 0.981 & 1 & --- \\
+\midrule
+\textit{Context}: independent noise    & --- & --- & --- & 0.778 & --- & --- \\
+\textit{Context}: 50 vs 200 steps      & --- & --- & --- & 0.996 & --- & --- \\
+\bottomrule
+\end{tabular}
+\end{table}

	@@ -0,0 +1,36 @@

+\begin{longtable}{llrrrrrr}
+\caption{Disease-effect agreement for each cell type in the Welch test. All models use the same donors and measured genes within each cell type. J counts shared genes and divides by all distinct genes in the two sets. Full J uses all genes called significant. The last column uses up to 50 significant genes ranked by absolute effect.}\label{tab:deg_by_type}\\
+\toprule
+Type & Checkpoint & $r$ & RMSE & Real DEGs & Gen. DEGs & Full $J$ & Top-50 $J$ \\
+\midrule
+\endfirsthead
+\toprule
+Type & Checkpoint & $r$ & RMSE & Real DEGs & Gen. DEGs & Full $J$ & Top-50 $J$ \\
+\midrule
+\endhead
+T cell & 2K E1 & 0.568 & 0.1095 & 835 & 770 & 0.396 & 0.299 \\
+T cell & 6K E1 & 0.528 & 0.1139 & 835 & 276 & 0.181 & 0.250 \\
+T cell & 11.6K E1 & 0.438 & 0.1195 & 835 & 273 & 0.159 & 0.235 \\
+T cell & 2K E2 & 0.648 & 0.1030 & 835 & 624 & 0.366 & 0.351 \\
+T cell & 2K E3 & 0.733 & 0.0969 & 835 & 427 & 0.304 & 0.389 \\
+\midrule
+B cell & 2K E1 & 0.373 & 0.1058 & 266 & 238 & 0.140 & 0.220 \\
+B cell & 6K E1 & 0.267 & 0.1116 & 266 & 30 & 0.042 & 0.067 \\
+B cell & 11.6K E1 & 0.145 & 0.1155 & 266 & 22 & 0.029 & 0.043 \\
+B cell & 2K E2 & 0.563 & 0.0952 & 266 & 215 & 0.206 & 0.316 \\
+B cell & 2K E3 & 0.660 & 0.0881 & 266 & 167 & 0.209 & 0.351 \\
+\midrule
+Monocyte & 2K E1 & 0.444 & 0.1340 & 289 & 261 & 0.211 & 0.149 \\
+Monocyte & 6K E1 & 0.244 & 0.1441 & 289 & 69 & 0.088 & 0.042 \\
+Monocyte & 11.6K E1 & 0.155 & 0.1500 & 289 & 95 & 0.082 & 0.042 \\
+Monocyte & 2K E2 & 0.649 & 0.1179 & 289 & 219 & 0.283 & 0.389 \\
+Monocyte & 2K E3 & 0.767 & 0.1010 & 289 & 279 & 0.318 & 0.562 \\
+\midrule
+NK cell & 2K E1 & 0.358 & 0.1370 & 231 & 131 & 0.168 & 0.205 \\
+NK cell & 6K E1 & 0.347 & 0.1385 & 231 & 58 & 0.095 & 0.176 \\
+NK cell & 11.6K E1 & 0.160 & 0.1509 & 231 & 85 & 0.097 & 0.099 \\
+NK cell & 2K E2 & 0.432 & 0.1344 & 231 & 95 & 0.164 & 0.282 \\
+NK cell & 2K E3 & 0.594 & 0.1207 & 231 & 110 & 0.209 & 0.333 \\
+\midrule
+\bottomrule
+\end{longtable}

	@@ -0,0 +1,17 @@

+\begin{table}[t]
+\centering
+\footnotesize
+\setlength{\tabcolsep}{3pt}
+\caption{Width effects, paired within schedule part. Each entry is the mean difference over parts 6--16 ($n=11$ paired parts) with its standard deviation, and the mean divided by its standard error. Positive values mean the wider model scores higher. Composition $L_1$ is lower-better, so a positive value there is worse. Two independently trained base models resolve differences of about 0.018 in $r$ and 7 in DEG overlap (Methods).}
+\label{tab:factorial}
+\begin{tabular}{llrrr}
+\toprule
+Change & Held fixed & $\Delta r$ & $\Delta$ DEG & $\Delta$ comp. \\
+\midrule
+Trunk $1\times\to4\times$ & cond.\ 512   & $-0.014\pm0.034$ ($-1.4$) & $-11.5\pm12.9$ ($-2.9$) & $-0.000\pm0.022$ ($-0.0$) \\
+Trunk $1\times\to4\times$ & cond.\ 2,048 & $+0.003\pm0.021$ ($+0.5$) & $-4.0\pm7.7$ ($-1.7$)   & $+0.114\pm0.034$ ($+11.2$) \\
+Cond.\ $512\to2{,}048$ & trunk $1\times$ & $-0.013\pm0.018$ ($-2.3$) & $-6.8\pm7.9$ ($-2.8$)   & $-0.027\pm0.024$ ($-3.7$) \\
+Cond.\ $512\to2{,}048$ & trunk $4\times$ & $+0.004\pm0.015$ ($+0.9$) & $+0.6\pm4.8$ ($+0.4$)   & $+0.087\pm0.039$ ($+7.3$) \\
+\bottomrule
+\end{tabular}
+\end{table}

	@@ -0,0 +1,18 @@

+\begin{table}[t]
+\centering
+\footnotesize
+\setlength{\tabcolsep}{4pt}
+\caption{Three ways to read the support head, on the same checkpoints. Base arm, parts 14--16 of the same run, same conditions, seed and 50 deterministic steps; only the last-step rule that turns support logits into a sparsity pattern differs. The inherited default keeps a gene when its detection probability exceeds one half. The swept threshold keeps it above $\mathrm{sigmoid}(-2.0)=0.12$, the setting that maximizes $r$ in a sweep (Supplementary Note~S5). The drawn read-out samples each gene from its own probability and has no parameter. Detected genes are means per cell on the 2,980-gene panel at part 16; the observed cells have 252 (median 230). Scores are means over the three parts.}
+\label{tab:readout}
+\begin{tabular}{lrrrrr}
+\toprule
+Read-out & parameters & genes/cell & $r$ & DEG & comp.\ $L_1$ \\
+\midrule
+Cut at $p>0.5$ (inherited default) & 1 & 191 & 0.245 & 59.7 & 0.199 \\
+Cut at $p>0.12$ ($\tau=-2.0$, swept) & 1 & 286 & \textbf{0.326} & 65.7 & 0.208 \\
+Drawn from $p$ (recommended)       & \textbf{0} & 228 & 0.286 & \textbf{67.7} & \textbf{0.197} \\
+\midrule
+Observed cells & --- & 252 & --- & --- & --- \\
+\bottomrule
+\end{tabular}
+\end{table}