Error with featureSubset = 'highQuality' and remove_uninformative_feature_outlier=TRUE
93 views
Skip to first unread message
Miguel Cosenza
Feb 15, 2021, 4:33:50 AM2/15/21
to MSstats
Dear MSstats team,
I am reporting and error that I am having while executing the dataProcess function:
I am reporting and error that I am having while executing the dataProcess function:
> normalized_data <- dataProcess(msts_formated_data,
+ logTrans=2,
+ normalization="equalizeMedians",
+ nameStandards=NULL,
+ address="",
+ fillIncompleteRows=TRUE,
+ featureSubset="highQuality",
+ remove_uninformative_feature_outlier=TRUE,
+ n_top_feature=3,
+ summaryMethod="TMP",
+ equalFeatureVar=TRUE,
+ censoredInt="NA",
+ cutoffCensored="minFeature",
+ MBimpute=FALSE,
+ remove50missing=FALSE,
+ maxQuantileforCensored=0.999,
+ clusters=NULL)
** Flag uninformative feature and outliers by feature selection algorithm.
Analyzing features in L channel...
Joining, by = c("protein", "feature")
Joining, by = "protein"
Joining, by = "protein"
Identifying low-coverage features...
Fitting robust linear models...
Joining, by = "protein"
Identifying outliers and calculating feature variances...
Error: Must specify either `data` or `newdata` argument.
In addition: Warning message:
10 very small variances detected, have been offset away from zero
Best wishes,
Miguel
> sessionInfo()
R version 4.0.3 (2020-10-10)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows Server x64 (build 17763)
Matrix products: default
locale:
[1] LC_COLLATE=English_United States.1252 LC_CTYPE=English_United States.1252
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C
[5] LC_TIME=English_United States.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] stringr_1.4.0 tidyr_1.1.2 readr_1.4.0 here_1.0.1 dplyr_1.0.4
[6] MSstats_3.22.0
loaded via a namespace (and not attached):
[1] tinytex_0.29 gtools_3.8.2 statmod_1.4.35
[4] minpack.lm_1.2-1 tidyselect_1.1.0 xfun_0.20
[7] reshape2_1.4.4 purrr_0.3.4 splines_4.0.3
[10] lattice_0.20-41 colorspace_2.0-0 vctrs_0.3.6
[13] generics_0.1.0 doSNOW_1.0.19 snow_0.4-3
[16] marray_1.68.0 survival_3.2-7 rlang_0.4.10
[19] pillar_1.4.7 nloptr_1.2.2.2 glue_1.4.2
[22] DBI_1.1.1 plyr_1.8.6 foreach_1.5.1
[25] lifecycle_0.2.0 munsell_0.5.0 gtable_0.3.0
[28] caTools_1.18.1 codetools_0.2-18 parallel_4.0.3
[31] preprocessCore_1.52.1 broom_0.7.4 Rcpp_1.0.6
[34] KernSmooth_2.23-18 scales_1.1.1 backports_1.2.1
[37] limma_3.46.0 lme4_1.1-26 gplots_3.1.1
[40] hms_1.0.0 ggplot2_3.3.3 stringi_1.5.3
[43] ggrepel_0.9.1 grid_4.0.3 rprojroot_2.0.2
[46] bitops_1.0-6 tools_4.0.3 magrittr_2.0.1
[49] tibble_3.0.6 crayon_1.4.1 pkgconfig_2.0.3
[52] MASS_7.3-53 ellipsis_0.3.1 Matrix_1.2-18
[55] data.table_1.13.6 assertthat_0.2.1 minqa_1.2.4
[58] iterators_1.0.13 R6_2.5.0 boot_1.3-25
[61] nlme_3.1-150 compiler_4.0.3
Many thanks in advance for taking a look.
Best wishes,
Miguel
Mateusz Staniak
Feb 15, 2021, 5:01:08 AM2/15/21
to MSstats
Hi,
thanks for letting us know about the error.
Please install the development version by running devtools::install_github("MeenaChoi/MSstats") and upgrade dependencies at the same time (you will need the devtools package for this).
If the problem still persists, please let me know. Then, a snippet of your data to reproduce the error would be helpful.
Kind regards,
Mateusz Staniak
Mateusz Staniak
Miguel Cosenza
Feb 15, 2021, 5:29:09 AM2/15/21
to MSstats
Hello Mateusz,
Many thanks for taking a look. The error is persistent after installing the development version.
I am sending a link with the data and the script to reproduce the error via direct e-mail.
Many thanks for taking a look. The error is persistent after installing the development version.
I am sending a link with the data and the script to reproduce the error via direct e-mail.
Best,
Miguel
Mateusz Staniak
Feb 15, 2021, 7:30:55 AM2/15/21
to MSstats
Hi,
currently, MSstats uses broom package to process output of fitting robust linear models while performing feature selection.
However, in case of two proteins in your datasets:
- sp|O15371|EIF3D_HUMAN
- sp|P37108|SRP14_HUMAN
the model is rank deficient, and broom package returns an error in this case.
We will fix this problem in the next MSstats release. I will send you results from this version (including these two proteins in a seperate e-mail).
With the current version, excluding these two proteins from your analysis should fix the problem.
Kind regards,
Mateusz
antje dittmann
Mar 10, 2021, 6:43:22 AM3/10/21
to MSstats
Hi Mateusz,
I'm seeing the same error:
> data_outlier_removed <- dataProcess(csv,
+ featureSubset = "highQuality",
+ remove_uninformative_feature_outlier = TRUE,
+ MBimpute = TRUE,
+ remove50missing = TRUE)
** Log2 intensities under cutoff = 17.618 were considered as censored missing values.
** Log2 intensities = NA were considered as censored missing values.
** Flag uninformative feature and outliers by feature selection algorithm.
Analyzing features in L channel...
Joining, by = c("protein", "feature")
Joining, by = "protein"
Joining, by = "protein"
Identifying low-coverage features...
Fitting robust linear models...
Joining, by = "protein"
Identifying outliers and calculating feature variances...
Fehler: Must specify either `data` or `newdata` argument.
Would it be possible to get a fix for this error in the development version already?
Kind regards,
Antje
Ben Polacco
May 19, 2021, 4:01:34 PM5/19/21
to MSstats
While the devs work on a real fix, I modified a forked copy that catches the error, reports the problem proteins, sets all features from problem proteins as non-outliers, and proceeds to finish the dataProcess. You can use the result as-computed, or remove the problem proteins and repeat dataProcess. To install my version:
library (devtools)
To test that it worked, look for the 9999 as a minor version:
> packageVersion("MSstats")
[1] ‘3.23.1.9999’
Reply all
Reply to author
Forward
0 new messages