LRScore calculation gives 'Number of valid intrs: 0 out of 754 from database.' and report error when running inferSignif()

[VivLon]

Hi,

When I ran inferIntrScore() and inferSignif() on my own dataset, it gave following messages:

== Calculating diffusible ligand-receptor scores ==

Computing LR-scores using diff_dep database.

- Ligand: GauEps, Receptor: Raw.

- Number of valid intrs: 0 out of 754 from database.

Warning message in max(as.integer(names(lig.fac))):
“no non-missing arguments to max; returning -Inf”
Permuting whole dataset 128 times...

- Permuting ligand Imp slot: GauEps...
- Permuting receptor Imp slot: Raw...


- Calculating NULL scores...

Permuting scores on Score slot: GauEps-Raw...

Warning message in max(as.integer(names(lig.fac))):
“no non-missing arguments to max; returning -Inf”
== Calculating contact-dependent ligand-receptor scores ==

Computing LR-scores using cont_dep database.

- Ligand: DT, Receptor: Raw.

- Number of valid intrs: 0 out of 109 from database.

Warning message in max(as.integer(names(lig.fac))):
“no non-missing arguments to max; returning -Inf”
Permuting whole dataset 128 times...

- Permuting ligand Imp slot: DT...


- Permuting receptor Imp slot: Raw...


- Calculating NULL scores...

Permuting scores on Score slot: DT-Raw...

Warning message in max(as.integer(names(lig.fac))):
“no non-missing arguments to max; returning -Inf”
Computing LR-scores using cont_dep database.

- Ligand: Raw, Receptor: Raw.

- Number of valid intrs: 0 out of 109 from database.

Warning message in max(as.integer(names(lig.fac))):
“no non-missing arguments to max; returning -Inf”
Permuting whole dataset 128 times...

- Permuting ligand Imp slot: Raw...


- Permuting receptor Imp slot: Raw...


- Calculating NULL scores...

Permuting scores on Score slot: Raw-Raw...

Warning message in max(as.integer(names(lig.fac))):
“no non-missing arguments to max; returning -Inf”
Inferring significant beads on Score slot GauEps-Raw... 

Error in dimnames(p.val.mtx) <- dimnames(i.mtx): 'dimnames' applied to non-array
Traceback:

1. inferSignif(cs, p.value = 0.05, reads.thresh = 100, sig.thresh = 100)
2. .inferSignif.matrix_like(object@counts, object@lrscore[[slot.use]]@score, 
 .     object@lrscore[[slot.use]]@score.null, fdr.method, nb.fac, 
 .     use.intr.db, [email protected][["gene_to_uniprot"]], p.value, 
 .     reads.thresh, sig.thresh)
3. getPvalues(lrscore.mtx, null.lrscore.mtx)

I guess it's because genes in my dataset doesn't overlap with the database you used. Do you have any suggestions?

[VivLon]

Just now, I tried using my entire dataset without filtering any genes, and the above problem doesn't appear this time. But rankIntrSpatialVar(cs) gives "Number of spatially variable interactions: 0 out of 0 high-quality interactions." Therefore, when I print the result of showIntr, it gives NA. Any suggestions on this?

[mvfki]

It looks like the problem is likely to be related to the function changeUniprot(). Did you skip this step in the tutorial or did you used some other customized database?

[VivLon]

No, I completely followed the tutorial and didn't skip this step or use other database.