EUCAST.RmdPCA.Rmdbenchmarks.Rmd
In the table above, all measurements are shown in milliseconds (thousands of seconds). A value of 5 milliseconds means it can determine 200 input values per second. It case of 100 milliseconds, this is only 10 input values per second.
To achieve this speed, the as.mo function also takes into account the prevalence of human pathogenic microorganisms. The downside of this is of course that less prevalent microorganisms will be determined less fast. See this example for the ID of Methanosarcina semesiae (B_MTHNSR_SEMS), a bug probably never found before in humans:
That takes 6.1 times as much time on average. We can conclude that looking up arbitrary codes of less prevalent microorganisms is the worst way to go, in terms of calculation performance. Full names (like Methanosarcina semesiae) are always very fast and only take some thousands of seconds to coerce - they are the most probable input from most data sets.
+# expr min lq mean median uq +# as.mo("metsem") 1485.000 1507.000 1524.00 1519.000 1538.000 +# as.mo("METSEM") 1371.000 1495.000 1557.00 1567.000 1633.000 +# as.mo("M. semesiae") 16.010 16.310 25.38 16.480 42.840 +# as.mo("M. semesiae") 15.700 15.900 16.74 16.370 17.480 +# as.mo("Methanosarcina semesiae") 5.885 6.116 11.79 6.347 8.155 +# max neval +# 1577.00 10 +# 1663.00 10 +# 48.53 10 +# 18.55 10 +# 32.92 10 +That takes 5.5 times as much time on average. We can conclude that looking up arbitrary codes of less prevalent microorganisms is the worst way to go, in terms of calculation performance. Full names (like Methanosarcina semesiae) are always very fast and only take some thousands of seconds to coerce - they are the most probable input from most data sets.
In the figure below, we compare Escherichia coli (which is very common) with Prevotella brevis (which is moderately common) and with Methanosarcina semesiae (which is uncommon):
Uncommon microorganisms take a lot more time than common microorganisms. To relieve this pitfall and further improve performance, two important calculations take almost no time at all: repetitive results and already precalculated results.
@@ -287,11 +272,11 @@ # keep only the unique ones unique() %>% # pick 50 of them at random - sample(50) %>% + sample(50) %>% # paste that 10,000 times rep(10000) %>% # scramble it - sample() + sample() # got indeed 50 times 10,000 = half a million? length(x) @@ -306,9 +291,9 @@ times = 100) print(run_it, unit = "ms", signif = 3) # Unit: milliseconds -# expr min lq mean median uq max neval -# mo_name(x) 564 605 673 630 657 1100 100 -So transforming 500,000 values (!!) of 50 unique values only takes 0.63 seconds (630 ms). You only lose time on your unique input values.
+# expr min lq mean median uq max neval +# mo_name(x) 542 585 605 601 614 738 100 +So transforming 500,000 values (!!) of 50 unique values only takes 0.6 seconds (600 ms). You only lose time on your unique input values.
So going from mo_name("Staphylococcus aureus") to "Staphylococcus aureus" takes 0.0009 seconds - it doesn’t even start calculating if the result would be the same as the expected resulting value. That goes for all helper functions:
So going from mo_name("Staphylococcus aureus") to "Staphylococcus aureus" takes 0.0008 seconds - it doesn’t even start calculating if the result would be the same as the expected resulting value. That goes for all helper functions:
run_it <- microbenchmark(A = mo_species("aureus"),
B = mo_genus("Staphylococcus"),
C = mo_name("Staphylococcus aureus"),
@@ -337,14 +322,14 @@
print(run_it, unit = "ms", signif = 3)
# Unit: milliseconds
# expr min lq mean median uq max neval
-# A 0.499 0.511 0.516 0.517 0.522 0.544 10
-# B 0.532 0.539 0.550 0.542 0.563 0.592 10
-# C 0.718 0.787 0.832 0.843 0.889 0.904 10
-# D 0.538 0.548 0.566 0.567 0.571 0.607 10
-# E 0.503 0.509 0.515 0.513 0.516 0.549 10
-# F 0.502 0.504 0.514 0.511 0.519 0.539 10
-# G 0.493 0.513 0.538 0.514 0.536 0.684 10
-# H 0.499 0.501 0.509 0.505 0.516 0.531 10Of course, when running mo_phylum("Firmicutes") the function has zero knowledge about the actual microorganism, namely S. aureus. But since the result would be "Firmicutes" anyway, there is no point in calculating the result. And because this package ‘knows’ all phyla of all known bacteria (according to the Catalogue of Life), it can just return the initial value immediately.
Currently supported are German, Dutch, Spanish, Italian, French and Portuguese.
diff --git a/docs/articles/benchmarks_files/figure-html/unnamed-chunk-4-1.png b/docs/articles/benchmarks_files/figure-html/unnamed-chunk-4-1.png index f86f4ef2..c7db6200 100644 Binary files a/docs/articles/benchmarks_files/figure-html/unnamed-chunk-4-1.png and b/docs/articles/benchmarks_files/figure-html/unnamed-chunk-4-1.png differ diff --git a/docs/articles/benchmarks_files/figure-html/unnamed-chunk-6-1.png b/docs/articles/benchmarks_files/figure-html/unnamed-chunk-6-1.png index 3eec84dd..4fa5b80c 100644 Binary files a/docs/articles/benchmarks_files/figure-html/unnamed-chunk-6-1.png and b/docs/articles/benchmarks_files/figure-html/unnamed-chunk-6-1.png differ diff --git a/docs/articles/index.html b/docs/articles/index.html index 28f4f619..db7952b8 100644 --- a/docs/articles/index.html +++ b/docs/articles/index.html @@ -78,7 +78,7 @@ diff --git a/docs/authors.html b/docs/authors.html index beceee87..1ef3c8a5 100644 --- a/docs/authors.html +++ b/docs/authors.html @@ -78,7 +78,7 @@ diff --git a/docs/countries.png b/docs/countries.png index e9e82ee6..62084e3f 100644 Binary files a/docs/countries.png and b/docs/countries.png differ diff --git a/docs/countries_large.png b/docs/countries_large.png index 24b75f2c..1756232f 100644 Binary files a/docs/countries_large.png and b/docs/countries_large.png differ diff --git a/docs/index.html b/docs/index.html index b33d8bda..6a5263fd 100644 --- a/docs/index.html +++ b/docs/index.html @@ -43,7 +43,7 @@ @@ -204,8 +204,8 @@ A methods paper about this package has been preprinted at bioRxiv (DOI: 10.1101/
- Used in almost 100 countries
- Since its first public release in early 2018, this package has been downloaded over 25,000 times from 99 countries (as of February 2020, CRAN logs). Click the map to enlarge.
This package is ready-to-use for specialists in many fields:
The latest and unpublished development version can be installed with (precaution: may be unstable):
-install.packages("devtools")
-devtools::install_gitlab("msberends/AMR")install.packages("remotes")
+remotes::install_gitlab("msberends/AMR")ggplot_pca() functionclean to cleaner, as this package was renamed accordingly upon CRAN requestas.mic())lintr packageFix for mo_shortname() where species would not be determined correctly
%like%, it will now return the callPercentages will now will rounded more logically (e.g. in freq function)
crayon, to support formatted text in the consoletidyr is now mandatory (went to Import field) since portion_df and count_df rely on itas.rsi and as.mic now add the package name and version as attributesREADME.md with more examplesGetting SNOMED codes of a microorganism, or get its name associated with a SNOMED code
Getting LOINC codes of an antibiotic, or get its name associated with a LOINC code
Machine reading the EUCAST and CLSI guidelines from 2011-2020 to translate MIC values and disk diffusion diameters to R/SI
Principal component analysis for AMR
Date of data entry
Date this data was entered in WHONET
AMP_ND10:CIP_EE
27 different antibiotics. You can lookup the abbreviatons in the antibiotics data set, or use e.g. ab_name("AMP") to get the official name immediately. Before analysis, you should transform this to a valid antibiotic class, using as.rsi().
data.frame with 452 observations and 14 variables:data.frame with 452 observations and 14 variables:ab
Antibiotic ID as used in this package (like AMC), using the official EARS-Net (European Antimicrobial Resistance Surveillance Network) codes where available
iv_units
Units of iv_ddd
An object of class data.frame with 102 rows and 9 columns.
World Health Organization (WHO) Collaborating Centre for Drug Statistics Methodology (WHOCC): https://www.whocc.no/atc_ddd_index/
diff --git a/docs/reference/example_isolates.html b/docs/reference/example_isolates.html index 062fef0d..b58598b2 100644 --- a/docs/reference/example_isolates.html +++ b/docs/reference/example_isolates.html @@ -79,7 +79,7 @@mo
ID of microorganism created with as.mo(), see also microorganisms
PEN:RIF
40 different antibiotics with class rsi (see as.rsi()); these column names occur in the antibiotics data set and can be translated with ab_name()
bacteria
info about microorganism that can be transformed with as.mo(), see also microorganisms
AMX:GEN
4 different antibiotics that have to be transformed with as.rsi()
code
Commonly used code of a microorganism
mo
ID of the microorganism in the microorganisms data set
prevalence
Prevalence of the microorganism, see as.mo()
snomed
SNOMED code of the microorganism. Use mo_snomed() to retrieve it quickly, see mo_property().
Catalogue of Life: Annual Checklist (public online taxonomic database), http://www.catalogueoflife.org (check included annual version with catalogue_of_life_version()).
ref
Author(s) and year of concerning scientific publication
prevalence
Prevalence of the microorganism, see as.mo()
Catalogue of Life: Annual Checklist (public online taxonomic database), http://www.catalogueoflife.org (check included annual version with catalogue_of_life_version()).
The pca() function takes a data.frame as input and performs the actual PCA with the R function prcomp().
The result of the pca() function is a prcomp object, with an additional attribute non_numeric_cols which is a vector with the column names of all columns that do not contain numeric values. These are probably the groups and labels, and will be used by ggplot_pca().
-The lifecycle of this function is experimental. An experimental function is in the very early stages of development. The unlying code might be changing frequently as we rapidly iterate and explore variations in search of the best fit. Experimental functions might be removed without deprecation, so you are generally best off waiting until a function is more mature before you use it in production code. Experimental functions will not be included in releases we submit to CRAN, since they have not yet matured enough.
+The lifecycle of this function is maturing. The unlying code of a maturing function has been roughed out, but finer details might still change. We will strive to maintain backward compatibility, but the function needs wider usage and more extensive testing in order to optimise the unlying code.
# `example_isolates` is a dataset available in the AMR package. @@ -334,7 +334,7 @@ The lifecycle of this function is experimenArguments Value Details -Experimental lifecycle +Maturing lifecycle Examples diff --git a/docs/reference/rsi_translation.html b/docs/reference/rsi_translation.html index b46ad1a2..4bae4fc2 100644 --- a/docs/reference/rsi_translation.html +++ b/docs/reference/rsi_translation.html @@ -79,7 +79,7 @@
breakpoint_S
Lowest MIC value or highest number of millimetres that leads to "S"
breakpoint_R
Highest MIC value or lowest number of millimetres that leads to "R"
The repository of this AMR package contains a file comprising this exact data set: https://gitlab.com/msberends/AMR/blob/master/data-raw/rsi_translation.txt. This file allows for machine reading EUCAST and CLSI guidelines, which is almost impossible with the Excel and PDF files distributed by EUCAST and CLSI. This file is updated automatically.
- Used in almost 100 countries
- Since its first public release in early 2018, this package has been downloaded over 25,000 times from 99 countries (as of February 2020, CRAN logs). Click the map to enlarge.