AttributeError: 'tuple' object has no attribute 'dim'

[pensive-bubble]

Hello again,

I have run into a further error while trying to run disease critical network demo. Tuple object causing further issues.

If you know how to fix this, it would be greatly appreciated!
Natasha

[kexinhuang12345]

Hi! are you using exactly the same code as in the demo? same model/gwas sumstats as well?

[pensive-bubble]

Hi! are you using exactly the same code as in the demo? same model/gwas sumstats as well?

Hello! I wasn't, I was using my own GWAS, but I've just tried it with the same model/gwas sumstats and have run into the same error.

[skoc]

Hi @pensive-bubble , can you make this small change and try again:

KGWAS/kgwas/utils.py

Line 461 in 68d6f60

x_dict = {i: j[0] for i,j in x_dict.items()}

That line will be:
x_dict = {i: j[0][0] for i,j in x_dict.items()}

[pensive-bubble]

Hi @pensive-bubble , can you make this small change and try again:

KGWAS/kgwas/utils.py

Line 461 in 68d6f60

x_dict = {i: j[0] for i,j in x_dict.items()}

That line will be:
x_dict = {i: j[0][0] for i,j in x_dict.items()}

Hello, I tried both on my GWAS and on the test one. I ended up with a fun new error.

[Himanshu-at-bayer]

Hi team, I am also having the same error while running the sample code :(

[skoc]

Hi @pensive-bubble , can you make this small change and try again:
KGWAS/kgwas/utils.py
Line 461 in 68d6f60
x_dict = {i: j[0] for i,j in x_dict.items()}
That line will be:
x_dict = {i: j[0][0] for i,j in x_dict.items()}

Hello, I tried both on my GWAS and on the test one. I ended up with a fun new error.

Thank you @pensive-bubble for the follow-up! Can make this pd concat change in your local version and try again:

# Instead of:
df_val_all = df_val_all.append(df_val)

# Use:
import pandas as pd
df_val_all = pd.concat([df_val_all, df_val], ignore_index=True)

Hey @Himanshu-at-bayer, you can also try x_dict = {i: j[0][0] for i,j in x_dict.items()} and pd concat change after that.

utils has multiple df_x.append(df_y) cases, changes has to be applied all

[Himanshu-at-bayer]

Thanks for the response and help. The first got removed after all the amendments you suggested but now got a new error

Retrieving weights...
Aggregating across node types...
Start generating disease critical network...
No filters... Using all genes and gene programs...

local/lib/python3.10/site-packages/kgwas/utils.py:590: FutureWarning: The provided callable <function std at 0x7f6dac644f70> is currently using SeriesGroupBy.std. In a future version of pandas, the provided callable will be used directly. To keep current behavior pass the string "std" instead.
rel2std = snp2genes_hit.groupby('rel_type').weight.agg(np.std).reset_index().rename(columns = {'weight': 'rel_type_std'})

ValueError Traceback (most recent call last)
Cell In[6], line 1
----> 1 df_network_weight, df_variant_interpretation, disease_critical_network = run.get_disease_critical_network(variant_threshold = 5e-8,
2 magma_path = None,
3 magma_threshold = 0.05,
4 program_threshold = 0.05,K_neighbors = 3,
5 num_cpus = 10)

File ~/.local/lib/python3.10/site-packages/kgwas/kgwas.py:272, in KGWAS.get_disease_critical_network(self, variant_threshold, magma_path, magma_threshold, program_threshold, K_neighbors, num_cpus)
268 def get_disease_critical_network(self, variant_threshold = 5e-8,
269 magma_path = None, magma_threshold = 0.05, program_threshold = 0.05,
270 K_neighbors = 3, num_cpus = 1):
271 df_network_weight = get_network_weight(self, self.data)
--> 272 df_variant_interpretation, disease_critical_network = generate_viz(self, df_network_weight, self.data_path, variant_threshold, magma_path, magma_threshold, program_threshold, K_neighbors, num_cpus)
273 return df_network_weight, df_variant_interpretation, disease_critical_network

File ~/.local/lib/python3.10/site-packages/kgwas/utils.py:600, in generate_viz(run, df_network, data_path, variant_threshold, magma_path, magma_threshold, program_threshold, K_neighbors, num_cpus)
598 v2g_hit = v2g_hit_with_rel_type[['h_idx', 't_idx', 'importance', 'h_type', 't_type', 'rel_type']]
599 v2g_hit.loc[:,'rel_type'] = v2g_hit.rel_type.apply(lambda x: x[4:])
--> 600 v2g_hit.loc[:,'Category'] = 'V2G'
602 v2g_hit.loc[:,'h_id'] = v2g_hit['h_idx'].apply(lambda x: idx2id['Gene'][x])
603 v2g_hit.loc[:,'t_id'] = v2g_hit['t_idx'].apply(lambda x: idx2id['SNP'][x])

File ~/.local/lib/python3.10/site-packages/pandas/core/indexing.py:911, in _LocationIndexer.setitem(self, key, value)
908 self._has_valid_setitem_indexer(key)
910 iloc = self if self.name == "iloc" else self.obj.iloc
--> 911 iloc._setitem_with_indexer(indexer, value, self.name)

File ~/.local/lib/python3.10/site-packages/pandas/core/indexing.py:1847, in _iLocIndexer._setitem_with_indexer(self, indexer, value, name)
1845 if not len(self.obj):
1846 if not is_list_like_indexer(value):
-> 1847 raise ValueError(
1848 "cannot set a frame with no "
1849 "defined index and a scalar"
1850 )
1851 self.obj[key] = value
1852 return

ValueError: cannot set a frame with no defined index and a scalar

[Himanshu-at-bayer]

This is line 590 in my case:

rel2std = snp2genes_hit.groupby('rel_type').weight.agg(np.std).reset_index().rename(columns = {'weight': 'rel_type_std'})

[jhchung]

I am also encountering this problem when there are no significant SNPs in the output. Some changes that seem to work (in addition to those already shared):

In the utils.py file, update lines 596, 613, and 631 to use assign instead of directly creating a new column.

For example, for line 596

# Old
v2g_hit.loc[:,'Category'] = 'V2G'

# Updated
v2g_hit = v2g_hit.assign(Category='V2G')

There is also a bug where the variant_threshold option in `get_disease_critical_network is not correctly used.

The GWAS P threshold is hardcoded in utils.py line 546

# Old 
hit_snps = gwas[gwas.P < 5e-8].SNP.values

# Updated
hit_snps = gwas[gwas.P < variant_threshold].SNP.values

[kexinhuang12345]

Nice! @jhchung would you like to make a PR for this? no worries if not, thanks!

[jhchung]

I need to do some additional testing to make sure these changes resolve the issue. but happy to make a PR when ready.

[kexinhuang12345]

Great - thank you!!

[Himanshu-at-bayer]

Thanks for the fix! Now at least I can run the codes.
However, I am not able to grasp the output of run.get_disease_crtiical_network :(
a. I can't reproduce the demo analysis even though I am running the same line of codes
b. df_variant_interpretation and disease_critical_network are returning empty even with low threshold cutoff for variant_threshold paramters.
c. How did you plotted the figures of variants connected with diseases/GO/other variants in the publication?