Error in if (child_partner1_mean_accuracy > child_partner2_mean_accuracy) { : missing value where TRUE/FALSE needed

[alifarhat40]

Hello,

I am running the following code on my seurat object:

rna_object
Active assay: RNA (32285 features, 0 variable features)
1 layer present: counts

rna_object <- NormalizeData(rna_object)
this correctly gives me the log normalized data.

Then I do this with the following paramters, which I believe is having issues being together:

rna_object <- CHOIR(rna_object ,
                        use_assay = "RNA",
                        feature_set = 'all',
                        batch_correction_method = "Harmony",
                        batch_labels = "sample",
                        n_var_features = 4000,
                        distance_awareness = FALSE,
                        use_variance = FALSE,
                        p_adjust = 'fdr',
                        alpha = 0.05,
                        downsampling_rate = 1,
                        n_cores= 100)

Here is the entire log output. The Error is the last sentence:

2025-04-20 02:08:06 AM : (Step 1/7) Checking inputs and preparing object..

Input data:
 - Object type: Seurat (v5)
 - # of cells: 32275
 - # of batches: 8
 - # of modalities: 1
 - ATAC data: FALSE
 - Countsplitting: FALSE
 - Assay: RNA
 - Layer used to build tree: data
 - Layer used to prune tree: data

Proceeding with the following parameters:
 - Intermediate data stored under key: CHOIR
 - Alpha: 0.05
 - Multiple comparison adjustment: fdr
 - Features to train RF: all
 - # of excluded features: 0
 - # of permutations: 100
 - # of RF trees: 50
 - Use variance: FALSE
 - Minimum accuracy: 0.5
 - Minimum connections: 1
 - Maximum repeated errors: 20
 - Distance approximation: TRUE
 - Maximum cells sampled: Inf
 - Downsampling rate: 1
 - Minimum reads: >0 reads
 - Maximum clusters: auto
 - Minimum cluster depth: 2000
 - Normalization method: none
 - Subtree dimensionality reductions: TRUE
 - Dimensionality reduction method: Default
 - Dimensionality reduction parameters provided: No
 - # of variable features: 4000
 - Batch correction method: Harmony
 - Batch correction parameters provided: No
 - Metadata column containing batch information: sample
 - Nearest neighbor parameters provided: 
     - verbose: FALSE
 - Clustering parameters provided: 
     - algorithm: 1
     - group.singletons: TRUE
     - verbose: FALSE
 - # of cores: 64
 - Random seed: 1

2025-04-20 02:08:06 AM : (Step 2/7) Running initial dimensionality reduction..
2025-04-20 02:08:06 AM : Preparing matrix using 'RNA' assay and 'data' slot..
2025-04-20 02:08:17 AM : Running PCA with 4000 variable features..
2025-04-20 02:08:50 AM : Running Harmony batch correction using column 'sample'..
2025-04-20 02:10:00 AM : (Step 3/7) Generating initial nearest neighbors graph..
2025-04-20 02:10:22 AM : (Step 4/7) Identify starting clustering resolution..
                      Starting resolution: 0.001
2025-04-20 02:11:14 AM : (Step 5/7) Building root clustering tree..

                      
                      Identified 2 clusters in root tree.
2025-04-20 02:12:55 AM : (Step 6/7) Subclustering root tree..
2025-04-20 02:14:55 AM : 10% (Subtree 1/2, 31877 cells), 2 total clusters.
2025-04-20 02:15:19 AM : 15% (Subtree 1/2, 31877 cells), 2 total clusters.
2025-04-20 01:08:16 PM : 35% (Subtree 1/2, 31877 cells), 102 total clusters.
2025-04-20 01:12:54 PM : 49% (Subtree 1/2, 31877 cells), 107 total clusters.
2025-04-20 01:28:29 PM : 64% (Subtree 1/2, 31877 cells), 113 total clusters.
2025-04-20 01:31:03 PM : 79% (Subtree 1/2, 31877 cells), 117 total clusters.
2025-04-20 01:33:18 PM : 94% (Subtree 1/2, 31877 cells), 119 total clusters.
2025-04-20 01:33:28 PM : 99% (Subtree 2/2, 398 cells), 119 total clusters.
2025-04-20 01:37:43 PM : 100% (Subtree 2/2, 398 cells), 126 total clusters.

2025-04-20 01:37:43 PM : (Step 7/7) Compiling full clustering tree..
                      Full tree has 69 levels and 124 clusters.

----------------------------------------
- CHOIR - Part 2: Prune clustering tree
----------------------------------------
2025-04-20 01:37:46 PM : (Step 1/2) Checking inputs and preparing object..

Input data:
 - Object type: Seurat
 - # of cells: 32275
 - # of batches: 8
 - # of modalities: 1
 - # of subtrees: 3
 - # of levels: 69
 - # of starting clusters: 124
 - Countsplitting: FALSE
 - Assay: RNA
 - Layer used to build tree: data
 - Layer used to prune tree: data

Proceeding with the following parameters:
 - Intermediate data stored under key: CHOIR
 - Alpha: 0.05
 - Multiple comparison adjustment: fdr
 - Features to train RF: all
 - # of excluded features: 0
 - # of permutations: 100
 - # of RF trees: 50
 - Use variance: FALSE
 - Minimum accuracy: 0.5
 - Minimum connections: 1
 - Maximum repeated errors: 20
 - Distance approximation: TRUE
 - Distance awareness: FALSE
 - All metrics collected: FALSE
 - Maximum cells sampled: Inf
 - Downsampling rate: 1
 - Minimum reads: >0 reads
 - Normalization method: none
 - Batch correction method: Harmony
 - Metadata column containing batch information: sample
 - Clustering parameters provided: 
     - algorithm: 1
     - group.singletons: TRUE
     - verbose: FALSE
 - # of cores: 64
 - Random seed: 1

2025-04-20 01:38:00 PM : (Step 2/2) Iterating through clustering tree..
Error in if (child_partner1_mean_accuracy > child_partner2_mean_accuracy) { : 
  missing value where TRUE/FALSE needed
Calls: CHOIR -> pruneTree
Execution halted

[cathrinesant]

Is this using the main or dev branch? If the main branch, would you try to see if this still occurs using the dev branch?

[alifarhat40]

This is on the main branch. Sure I will try to do it on the dev branch.

[alifarhat40]

Now I get a different error on the dev branch:

Error in .validInput(distance_awareness, "distance_awareness") : 
  Input value for 'distance_awareness' must be single value of class 'numeric', please supply valid input!
Calls: CHOIR -> .validInput
Execution halted

Apparently distance_awareness cannot be false here even though in your documentation you say it can be?

https://www.choirclustering.com/reference/CHOIR.html#:~:text=To%20omit%20all%20distance%20calculations%20and%20perform%20permutation%20testing%20on%20all%20comparisons%2C%20set%20this%20parameter%20to%20FALSE.%20Setting%20this%20parameter%20to%20FALSE%20or%20increasing%20the%20input%20value%20will%20increase%20the%20number%20of%20permutation%20test%20comparisons%20run%20and%2C%20thus%2C%20the%20computational%20time.

[cathrinesant]

In the dev branch, distance_awareness should be set to Inf to disable it.

[alifarhat40]

@catpetersen When setting it to Inf it gives:

Error in if (child_partner1_mean_accuracy > child_partner2_mean_accuracy) { : 
  missing value where TRUE/FALSE needed
Calls: CHOIR -> pruneTree
Execution halted