Add new element

[uby76]

Thank you for developing such a nice package, it is really convenient for us workers who are more familiar with R language!

I have a question, when my data if there are still Cl or Br elements, when calling as.peakData, how should I add to peakObj = as.peakData(e_data = data_fticrms, f_data = emeta, e_meta = e_fticrms, edata_ cname = “Mass”, mass_cname = “Mass”, fdata_cname = “SampleID”, c_cname = “C”, h_cname = “H”, o_cname = “O”, n_cname = “N”, s_cname = “S”, p_cname = “P”). Because I realized that there is nothing about Cl and Br inside the default parameters.

[clabornd]

If you have installed from the latest release (1.1.0) the object creation only handles CHNOSP. We have implemented support for arbitrary elements in #25 but have not done a release with these changes. You can still install these changes by installing from the latest master: remotes::install_github("EMSL-Computing/ftmsRanalysis").

In the new framework, we specify element columns with a list like so:

peakObj <- as.peakData(ftms12T_edata, ftms12T_fdata, ftms12T_emeta, 
                       edata_cname="Mass", fdata_cname="SampleID", 
                       mass_cname="Mass", element_col_names = list("C"="C", "H"="H", "O"="O", "N"="N", "S"="S", "P"="P"), 
                       isotopic_cname = "C13", 
                       isotopic_notation = "1")

Where the format of element_col_names is list("element_name" = "corresponding_count_column", ...). You can see this in the ftmsRanalysis.Rmd vignette. In your case you should be able to specify element_col_names = list("C"="C", "H"="H", "Cl" = "your_cl_count_column", "Br" = "your_br_count_column")

Currently there is limited functionality for these non-CHNOSP elements: you can just calculate element ratios with these new elements (see ?calc_element_ratios for details). Feel free to put in a feature request for any downstream processing of these new elements. Glad you have found the package useful!

[uby76]

Thank you for your quick response! Your answer is incredibly helpful for my analysis. I will update to the latest version and test it out. This package is indeed very convenient, and I will continue using it. Many thanks!

I have another question, I see that you used G-Test in ftmsRanalysis Introduction - ftmsRanalysis at the end when comparing the difference between two compositions, if my sample size is not that large. I mean, for one treatment I have only one sample, so how should I solve it?

[clabornd]

No problem, for your issue with the single sample, you will have to exclude this sample from any statistical tests.

group_designation should automatically set these group levels to NA. Then when calling this code in the vignette:

byGroup <- divideByGroupComparisons(peakObj, 
                                comparisons = "all")[[1]]$value

crop_unique <- summarizeGroupComparisons(byGroup, 
            summary_functions="uniqueness_gtest", 
            summary_function_params=list(
                  uniqueness_gtest=list(pres_fn="nsamps", 
                          pres_thresh=2, pvalue_thresh=0.05)))

head(crop_unique$e_data)

You could specify something like comparisons = list(c("A", "B")) in divideByGroupComparisons where "A" and "B" are groups with more than 1 sample. If you run into issues, you can also use subset.peakData to remove your singleton group and perform analysis on the remaining data.

[uby76]

Thank you so much for your response, this is really cool!