specific gene trends

[wangjiawen2013]

Hi,
Are there any ways to plot gene expression along only one or more lineages using via.get_gene_expression?
It's similar to palantir function compute_gene_trends(pr_res, gene_exprs, lineages=None, n_jobs=-1), where lineages: Subset of lineages for which to compute the trends

[ShobiStassen]

Hi again, Yes we can do that, I'll make the changes by Monday and update the git and pip versions

…

On Sat, 12 Nov 2022, 13:31 jiawen wang, ***@***.***> wrote: Hi, Are there any ways to plot gene expression along only one or more lineages using via.get_gene_expression? It's similar to palantir function compute_gene_trends(pr_res, gene_exprs, lineages=None, n_jobs=-1), where lineages: Subset of lineages for which to compute the trends — Reply to this email directly, view it on GitHub <#25>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AISI4SEOKRFH6K5W2C64JADWH4TT5ANCNFSM6AAAAAAR6E2SZE> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[ShobiStassen]

hi, just made the changes and are available on git/pip if you want to update. let me know if it works

[wangjiawen2013]

Great, it works!
And can we set a list of user defined terminal states ? Now the terminal states are generated automatically.

This is what palantir does:
terminal_states = pd.Series(['DC', 'Mono', 'Ery'],index=['Run5_131097901611291', 'Run5_134936662236454', 'Run4_200562869397916'])
start_cell = 'Run5_164698952452459'
pr_res = palantir.core.run_palantir(ms_data, start_cell, num_waypoints=500, terminal_states=terminal_states.index)

[ShobiStassen]

hi, that would be useful if you want to 'overwrite' the autodetection. we can do a fairly quick fix on that but would need until next week when we have a little more time. thanks for the useful suggestions

[ShobiStassen]

hi @wangjiawen2013 this should be fixed now - let me know if it works for you
you can specify user_defined_terminal_group=['pDC','MONO1','MEGA1'] - these must correspond to labels in true_label paramter, or you can give user_defined_terminal_cell=[] (with single cell indices as a list of integers) when you initialize v0=VIA()

you can also selectively plot lineage branch probabilities in draw_sc_lineage_probability() or get_gene_expression() by setting marker_lineages = [] with the suitable terminal lineage clusters.
draw_sc_lineage_probability(v0, embedding=tsnem, marker_lineages=[2,4,6,9]) get_gene_expression(via0=v0, gene_exp=df_magic, cmap='rainbow', marker_genes=marker_genes,marker_lineages=[2,4,6,9])
Randomseed is also now fixed to keep pt stable

[wangjiawen2013]

Hi,
Good feature !
will this feature introduce a bug when there are only terminal state ? Becuase I met this error:
In [57]: via.draw_sc_lineage_probability(via_coarse=v0, via_fine=v0,embedding=embedding,scatter_size=size)
2022-12-19 19:25:32.144658 Marker_lineages: [6]
2022-12-19 19:25:32.157409 The number of components in the original full graph is 1
2022-12-19 19:25:32.157447 For downstream visualization purposes we are also constructing a low knn-graph
2022-12-19 19:25:33.445752 Check sc pb [1.]
/home/wangjw/programs/miniconda3/envs/py37/lib/python3.7/site-packages/pyVIA/plotting_via.py:500: RuntimeWarning: Couldn't reach some vertices at src/paths/unweighted.c:440
path_orange = G_orange.get_shortest_paths(via_fine.root[ii], to=fst_i)[0]

TypeError Traceback (most recent call last)
in
----> 1 via.draw_sc_lineage_probability(via_coarse=v0, via_fine=v0,embedding=embedding,scatter_size=size)

~/programs/miniconda3/envs/py37/lib/python3.7/site-packages/pyVIA/plotting_via.py in draw_sc_lineage_probability(via_coarse, via_fine, embedding, idx, cmap_name, dpi, scatter_size, marker_lineages, fontsize)
524
525 if fig_nrows ==1:
--> 526 plot_sc_pb(axs[c], fig, embedding, p1_sc_bp[:, loc_ts_current], ti= str(ts_current)+'-'+str(majority_composition), cmap_name=cmap_name, scatter_size=scatter_size, fontsize=fontsize)
527
528 else:

TypeError: 'AxesSubplot' object is not subscriptable

[wangjiawen2013]

Perhaps all the function that use plt.subplots() will lead to this issue when there are only one plot.

[ShobiStassen]

hi, i fixed this bug in draw_sc_lineage_probability() so that you can also plot just one single lineage