Error in palimpsest_dfPosXSegm() for small data sets

[AndreVidas]

Hi Palimpsest.

I'm currently running Palimpsest on some downsampled data. I have 10 replicates of each downsampled size category and in a few instances of these replicas I get some errors in what seems to be a palimpsest_dfPosXSegm() call from the preprocessInput_sv() function:

Warning messages:
1: replacing previous import 'colorspace::plot' by 'graphics::plot' when loading 'NMF'
2: replacing previous import 'dendextend::cutree' by 'stats::cutree' when loading 'NMF'
3: replacing previous import 'NMF::entropy' by 'lsa::entropy' when loading 'Palimpsest'
4: replacing previous import 'NMF::dispersion' by 'plotrix::dispersion' when loading 'Palimpsest'
5: replacing previous import 'gplots::plotCI' by 'plotrix::plotCI' when loading 'Palimpsest'
Error in [<-.data.frame(*tmp*, , namesColsToAdd, value = c(NA_character_, :
replacement has 2859 rows, data has 4
Calls: preprocessInput_sv -> palimpsest_dfPosXSegm -> [<- -> [<-.data.frame
Execution halted

All the error messages I get resembles each other but with different amount of "X rows, data has X".

Do you have any idea of why this happens and if it's possible to circumvent this with this low amount of data? Or is it maybe just a consequence of too little data input?

The data size where these errors occur is in the range 8-64 samples and 198-5070 SV calls.
Best,
André

[jayendrashinde91]

Hi André,
Could you kindly be able to send me a reproducible example of the data you are using which generates this error? It will help me look at the same error and to better understand it.
Thanks,
Jay

[AndreVidas]

Sure.
Here I have attached three downsampled data sets each containing 8 randomly chosen samples with SV calls:

downsampled_sample-wise_8_rep_1_TEST.csv - should crash before the "Annotating mutation data:"-process bar has finished.

downsampled_sample-wise_8_rep_2_TEST.csv - should crash right after the "Annotating mutation data:"-process bar has finished.

downsampled_sample-wise_8_rep_3_TEST.csv - should complete without crash (for comparison)

Best,
André

testData.zip

[jayendrashinde91]

Dear André,
Thank you for providing these specific example data, makes life very easy while debugging.
As you mentioned, your downsampled data had low amount of data. This led to some difficulty while annotating the positions with the genes on reference genome. I have pushed the latest patch in the code for this issue which should run smoothly for any of your data now. Please try reinstalling Palimpsest and running the analysis for your data.
Thanks again for providing us with informative issues in the package.
Best wishes,

Jay

[AndreVidas]

Thanks so much Jay.

Now it works for the small data sets.

In one of the small data sets I now instead get one warning from Palimpsest describing that there were some errors in the NMF function:

Warning messages:
1: replacing previous import 'colorspace::plot' by 'graphics::plot' when loading 'NMF'
2: replacing previous import 'dendextend::cutree' by 'stats::cutree' when loading 'NMF'
3: replacing previous import 'NMF::entropy' by 'lsa::entropy' when loading 'Palimpsest'
4: replacing previous import 'NMF::dispersion' by 'plotrix::dispersion' when loading 'Palimpsest'
5: replacing previous import 'gplots::plotCI' by 'plotrix::plotCI' when loading 'Palimpsest'
Timing stopped at: 154.8 27.7 190.4
Timing stopped at: 154.4 27.2 189.6
Timing stopped at: 0.065 0 1.739
Timing stopped at: 154.8 27.18 190
Timing stopped at: 0.046 0 1.197
Timing stopped at: 107.8 19.16 224.2
Warning messages:
1: In (function (...) :
NAs were produced due to errors in some of the runs:
-#4[r=5] -> NMF::nmf - 1/100 fit(s) threw an error.

Error(s) thrown:

NA
-#5[r=6] -> NMF::nmf - 1/100 fit(s) threw an error.

Error(s) thrown:

NA
-#6[r=7] -> NMF::nmf - 1/100 fit(s) threw an error.

Error(s) thrown:

NA
-#8[r=9] -> NMF::nmf - 1/100 fit(s) threw an error.

Error(s) thrown:

NA
-#9[r=10] -> elements of 'k' must be between 1 and 8 [in call to 'stats::cutree']
-#10[r=11] -> elements of 'k' must be between 1 and 8 [in call to 'stats::cutree']
-#11[r=12] -> elements of 'k' must be between 1 and 8 [in call to 'stats::cutree']
2: Removed 20 rows containing missing values (geom_path).
3: Removed 72 rows containing missing values (geom_point).
Warning messages:
1: In sig.fit$res.x/sig.tot :
Recycling array of length 1 in vector-array arithmetic is deprecated.
Use c() or as.vector() instead.

2: In sig.fit$res.x/sig.tot :
Recycling array of length 1 in vector-array arithmetic is deprecated.
Use c() or as.vector() instead.

3: In sig.fit$res.x/sig.tot :
Recycling array of length 1 in vector-array arithmetic is deprecated.
Use c() or as.vector() instead.

4: In sig.fit$res.x/sig.tot :
Recycling array of length 1 in vector-array arithmetic is deprecated.
Use c() or as.vector() instead.

5: In sig.fit$res.x/sig.tot :
Recycling array of length 1 in vector-array arithmetic is deprecated.
Use c() or as.vector() instead.

6: In sig.fit$res.x/sig.tot :
Recycling array of length 1 in vector-array arithmetic is deprecated.
Use c() or as.vector() instead.

7: In sig.fit$res.x/sig.tot :
Recycling array of length 1 in vector-array arithmetic is deprecated.
Use c() or as.vector() instead.

8: In sig.fit$res.x/sig.tot :
Recycling array of length 1 in vector-array arithmetic is deprecated.
Use c() or as.vector() instead.

However I still get result output that seems to be fine. Nonetheless I have attached the data so that you can try it out yourself. Please tell me if you know that something is wrong with this run (i.e. the results can not be trusted).

Best,
André

downsampled_sample-wise_8_rep_8_TEST.csv.zip