|
{"metadata":{"id":"00370b93caa6c92362c84637fd6b62f1","source":"gardian_index","url":"https://cgspace.cgiar.org/rest/bitstreams/4bbf6e58-242e-42e2-983d-020908895ad5/retrieve"},"pageCount":5,"title":"","keywords":[],"chapters":[{"head":"","index":1,"paragraphs":[{"index":1,"size":24,"text":"Applications in Plant Sciences 2013 1 ( 6 ): 1300024; http://www.bioone.org/loi/apps © 2013 Simon and Spooner . Published by the Botanical Society of America."},{"index":2,"size":16,"text":"This article is a U.S. Government work and is in the public domain in the USA."}]},{"head":"Ap Applications ons in in Pl Plant t Scien Sciences ces","index":2,"paragraphs":[{"index":1,"size":123,"text":"A tedious part of preparing a taxonomic monograph is collating specimen data for (1) specimen citations and (2) indices to associate collectors and collector numbers to taxonomic identifi cations ( exsiccatae ). While preparing a taxonomic monograph of wild potatoes for northern South America, we designed tools to effi ciently and accurately convert specimen data into the formats specifi ed by Systematic Botany Monographs . One of the tools is a conversion tool to reformat tabulated records into specimen citations and the other for exsiccatae. It is a specifi c implementation of the reproducible research approach ( Gentleman and Temple Lang, 2004 ). A version of our software tool is freely available for R ( R Core Team, 2012 ) at http://cran.r-project.org/ web/packages/exsic/."}]},{"head":"METHODS AND RESULTS","index":3,"paragraphs":[{"index":1,"size":166,"text":"Overall design of the package -The package was designed to facilitate very specifi c steps in the processing of tabular specimen data: the conversion from table-oriented to index-based or list-based formats. Other necessary steps, such as data cleansing or data conversions, are not part of the main functions as there are other tools available for that purpose. That is, date conversions or geographical coordinate transformations into the fi nal desired format must be done separately by the user and before using the exsic function. The main function assumes, therefore, correct structure and content, but has been designed to be robust and to gracefully handle missing columns or content. In the case of essential columns (for sorting or fi ltering, e.g., \"species\", \"country\", \"collcite\", \"number\", \"majorarea\", \"minorarea\"), these will be added and fi lled with meaningful values for missing content (e.g., \"Anonymous\" for missing collector information). However, a few convenience functions have been added to handle specifi c transformations, and these are explained in more detail below."},{"index":2,"size":300,"text":"Tabular data format -As a starting point for preparing the specimen data, we used a format based on conventions defi ned by the BRAHMS (Botanical Research and Herbarium Management Software) package ( Filer, 2001 ), as detailed in an online manual ( Filer, 2010 ). The column names used by exsic are mostly the same as those used by BRAHMS, with a few modifi cations and additions to facilitate both the concurrent use of the BRAHMS conventions and this package ( Table 1 ) . One is the addition of the \"colldate\" fi eld for the \"date of collection\". The content of this fi eld should usually have the format \"1 Jan 2013\"; the corresponding BRAHMS fi eld names are: \"colldd\", \"collmm\", and \"collyy\" for day, month, and year, respectively. Functions in the package \"date\" can be used to convert from these three columns to the desired format and stored in the column \"colldate\". The other addition is the column or fi eld \"collcite\". The information on citing collectors is stored in two BRAHMS fi elds: \"collector\" and \"addcoll\". A custom function (\"coll.cite\") is provided that helps address most formatting issues, including whether or not to use initials, whether or not to use periods after initials, and whether to cite both names for collector pairs or cite only the fi rst and add \"et al.\" when there are more than two. The variations can be set using \"coll.cite\" function parameters. Another convenience function is called \"strip.last.dot\"; this is meant to process the column \"locnotes\" for location notes because they are separated by a comma in most specimen citations. A last variation is the use of a separate column for \"phenology\", with expected content to be \"fl \" for fl owering stage, \"fr\" for fruiting stage, or empty for neither stage."},{"index":3,"size":291,"text":"Target formats for specimen citations of numbered collections -We researched format variations using a Google search and found three additional formats to that used in Systematic Botany Monographs . The four formats ( Systematic Botany Monographs [format.SBMG], American Society of Plant Taxonomists [format.ASPT], New York Botanical Garden [format.NYBG], and PhytoKeys [format.PK]) are summarized in Table 2 . The main differences between these specimen citation formats relate to the formatting of species and country information as well as to differences in formatting the collector information. New York Botanical Garden journals include phenology information after the date of observation. Numbered collections formats differed with respect to using index numbers to point to the species or directly using a species. Another difference is whether consecutive numbers of the same species for a 1 Manuscript received 28 March 2013; revision accepted 22 May 2013. This research was supported by the National Science Foundation (DEB 0316614). The authors thank two anonymous reviewers for their feedback on earlier versions, which greatly helped to improve the scope and usability of the tool. 4 Author for correspondence: r. • Premise of the study: Taxonomists manage large amounts of specimen data. This is usually initiated in spreadsheets and then converted for publication into locality lists and indices to associate collectors and collector numbers from herbarium sheets to identifi cations ( exsiccatae ). This conversion process is mostly done by hand and is time-consuming, cumbersome, and errorprone. • Methods and Results: We constructed a tool, 'exsic,' based on the statistical software R. The exsic function is part of the R package 'exsic' and produces specimen citations and exsiccatae conforming to four related formats. • Conclusions: The tool increases speed, effi ciency, and accuracy to convert raw spreadsheet tables to publication-ready content."},{"index":4,"size":8,"text":"Key words: exsic; exsiccatae; R; reproducible research; software."}]},{"head":"Usage -","index":4,"paragraphs":[{"index":1,"size":263,"text":"The package can be downloaded from the central R repository CRAN using standard procedures in R and activated using the command \"library(exsic)\". A user would typically start using the main exsic function and either provide an inmemory table or a fi le path. For example: a table may be read using the custom function \"data <-read.exsic(fi lepath)\". This function ensures that all obligatory fi elds (see Table 1 ) are present; if not, they will be created and prefi lled with placeholder text (e.g., missing collector citation in \"collcite\" will be replaced with \"Anonymous\"; missing collection numbers with \"s.n.\"; missing dates with \"s.d.\"; and missing minor area or major area information with \"Unknown major/minor area\"). An example table with 1000 records of wild potato specimens is included in the package and can be accessed using \"system.fi le(\"samples/exsic.csv\", packages=\"exsic\")\". This may take a minute to process. For quicker testing, a subset can be constructed using \"pt = potato[1:10,]\" which will use only the fi rst 10 records. A complete minimal example using the provided sample table is given here: afi le <-system.fi le(\"samples/exsic.csv\", package=\"exsic\") potato <-read.exsic(afi le) pt = potato[1:10,] exsic(pt) # or shorthand for the whole table: exsic(fi le=\"afi le\") collector citation are grouped or not. The report created by the command example (exsic) creates a Web page as shown in Fig. 1 . The HTML page can then be further edited and formatted using a word processor. A user's guide showing the principal usage along with a precise description of the input format can be found in the package documentation (http://cran.r-project.org/web/packages/ exsic/exsic.pdf)."}]},{"head":"Sorting and Filtering -","index":5,"paragraphs":[{"index":1,"size":348,"text":"The specimen citation records are, by default, not fi ltered from the given table, and species and countries are sorted alphabetically. However, a taxonomist may want to order species by taxonomic relationships, and countries may be ordered from north to south and west to east. Also, the same table may be used for checking subsets. Therefore, to allow both fi ltering and custom sorting \"on-the-fl y,\" an additional table (an R data. frame) can be defi ned. It has two columns, \"country\" and \"species\", where the desired countries and species may be listed separated by semicolons without spaces. To list all countries or all species, the word \"all\" may be used. The table may be added to the exsic function as a parameter \"sortfi lter\". If this parameter is omitted, default options apply. A helper function to check format compliance is available (is.sortfi lter). A sample \"sortfi lter\" table is also available (sort.specs). c The \"phenology\" fi eld is only used in one format. http://www.bioone.org/loi/apps or species must be written exactly as in the primary table and must be separated by semicolons without spaces. Only those countries or species recognized will be used in the fi nal indices and will be sorted according to the provided sequence. C. Advanced users may defi ne their own format conventions using one of the \"format.XXXX\" (where XXXX is either SBMG, ASPT, NTBG, or PK) examples as a starting point. The formats are defi ned in simple tables so that they can be edited using spreadsheet software. The available formats for text include: bold, italic, underline or underscore (no difference), capitals, uppercase, parentheses, or square brackets. These formats can be combined and are applied from left to right; several options must be separated by a semicolon without a space. Unrecognized formatting words are ignored. For example, the formatting \"();italics\" will result in italicized parentheses whereas \"italics;()\" will not. More details can be found in Table 3 . D. The parameter \"header\" can be used to set other section titles; the parameter \"out.fi le\" is used to set the output fi le name."},{"index":2,"size":20,"text":"The resulting Web page will be in the same directory called \"exsic.html\" (this name can be changed via a parameter)."},{"index":3,"size":106,"text":"By default, the whole table passed to exsic will be used, sorted alphabetically by species and country, and formatted according to the conventions of Systematic Botany Monographs . To see an example, one may also type \"example(exsic)\" in the R console. The user may modify the defaults by combinations of the following fi ve options: (a) choose of one of four formats, (b) select a subset of species and countries, and order nonalphabetically by species or country, (c) defi ne additional formats, (d) set section titles and the output fi le name, and (e) selectively execute only some of the subroutines. These options are explained below."},{"index":4,"size":192,"text":"A. Choice of formats: The main exsic function has an additional parameter, \"formats\", that accepts tabulated parameters. The table must have only four columns named exactly as \"fi eld\", \"style\", \"sept\", \"comments\" (for example contents see Table 2 ). The four formats are listed in Table 2 . The default format is \"format.SBMG\". B. The \"sortfi lter\" is a parameter in the form of a table that must have only two columns, named \"country\" and \"species\". Within each fi eld, countries a The \"rec\" variable is the fi rst record from the table. It can be created using: rec = read.exsic(system.fi le(\"samples/exsic.csv\", package=\"exsic\"))[1,] http://www.bioone.org/loi/apps E. The building blocks of the main exsic function are also available individually along with some helper functions. Each index function can be executed separately. Output format is in an intermediate format called \"markdown\" format and needs to be converted. This can be achieved using the function \"write.exsic\". Headers can be provided by using \"exsic.header\" and a combination of string concatenating commands such as \"paste\". For example, a custom index could be achieved using: hdr = exsic.header(\"A header\") idx = index.collections(pt) txt = paste(hdr, idx, sep=\"\") write.exsic(txt, \"idx.html\")"},{"index":5,"size":7,"text":"For more details, see the package documentation."},{"index":6,"size":132,"text":"Implementation and speed -We used R and two libraries: (1) stringr ( Wickham, 2012 ) and (2) markdown package ( Gruber and Swartz, 2004 ;Allaire et al., 2012 ) as a basis to implement the exsic package. The exsic package shows one application of R and reproducible research tools for taxonomists and botanists. The package is freely available under the open source GNU Public License (GPL) and for all platforms supported by R (currently Windows, Linux, and Mac OS). On a Dell T7400 precision PC with 4 GB of RAM, a 2.66-GHz Intel Xeon CPU E5430 processor, and running R 2.15.2 on top of Ubuntu 12.10, the sample table with 1000 records was processed in about 30 s. The function has also been tested on Windows XP and Mac OS Snow Leopard."},{"index":7,"size":60,"text":"Important note-When working across operating systems with tables created in Excel, it is indispensable to make sure that data are saved not only in .csv format but also using the encoding standard UTF-8. This allows the use of accents or other alphabets in the indices. Excel does not use UTF-8 as a default, and this will result in formatting errors."}]},{"head":"CONCLUSIONS","index":6,"paragraphs":[{"index":1,"size":15,"text":"The tool primarily increases the speed of preparing specimen citations, numbered collections, and supporting indices."},{"index":2,"size":102,"text":"It also minimizes human errors in manual transcription from table to list formats as well as formatting errors. The possibility to quickly create indices in familiar formats also provides an opportunity to double check the consistency and completeness of the table before fi nal publication, thereby increasing the fi nal quality of the table and interpretation. We are not aware of a similar freely available tool except for the report module in the BRAHMS software version 7 that also facilitates the generation of these indices. While the BRAHMS software is Windows only, the exsic package works on Windows, Linux, and Mac OS."}]},{"head":"LITERATURE CITED","index":7,"paragraphs":[]}],"figures":[{"text":"Fig. 1 . Fig. 1. An example result for the exsic function using fi ve records from a sample table. "},{"text":" Integrated IT and Computational Research Unit, International Potato Center, Avenida La Molina 1895, La Molina, Lima, Peru; and 3 USDA-Agricultural Research Service, Vegetable Crops Research Unit, University of Wisconsin, 1575 Linden Drive, Madison, Wisconsin 53706-1590 USA CREATING INDICES 1 CREATING INDICES 1 REINHARD SIMON 2,4 AND DAVID M. SPOONER 3 REINHARD SIMON 2,4 AND DAVID M. SPOONER 3 "},{"text":"TABLE 1 . The data dictionary for the primary table of specimens. Field or column names correspond largely to BRAHMS standard; sample content data are invented. Exsic column name BRAHMS fi eld name Exsic required Description Type Sample content Exsic column nameBRAHMS fi eld nameExsic requiredDescriptionTypeSample content id NA Obligatory Continuous number from 1 to n Integer 1 idNAObligatoryContinuous number from 1 to nInteger1 genus genus Obligatory Genus Text Solanum genusgenusObligatoryGenusTextSolanum species sp1 Obligatory Species Text tuberosum speciessp1ObligatorySpeciesTexttuberosum collector collector Obligatory Collector name Text Linne, C collectorcollectorObligatoryCollector nameTextLinne, C number number Obligatory Collector number identifying the specimen Integer 1111 numbernumberObligatoryCollector number identifying the specimenInteger1111 addcoll addcoll Recommended Additional collector names Text Author, A addcolladdcollRecommendedAdditional collector namesTextAuthor, A collcite NA Obligatory Final citation of collector(s) Text Author collciteNAObligatoryFinal citation of collector(s)TextAuthor dups dups Recommended Duplicated herbaria Text PAR dupsdupsRecommendedDuplicated herbariaTextPAR majorarea majorarea Recommended Major subnational level Text Puno majorareamajorareaRecommendedMajor subnational levelTextPuno minorarea minorarea Recommended Second subnational level Text Puno minorareaminorareaRecommendedSecond subnational levelTextPuno locnotes locnotes Recommended Location Text at the shore of lake Titicaca locnoteslocnotesRecommendedLocationTextat the shore of lake Titicaca altitude alt Optional Elevation in meters Text 4000 m altitudealtOptionalElevation in metersText4000 m latitude lat Optional Latitude Text 12 ° 1 ′ 23 ″ N latitudelatOptionalLatitudeText12 ° 1 ′ 23 ″ N longitude long Optional Longitude Text 12 ° 1 ′ 23 ″ E longitudelongOptionalLongitudeText12 ° 1 ′ 23 ″ E colldate NA Recommended Collection date Text 1 Jan 2013 colldateNARecommendedCollection dateText1 Jan 2013 country country Obligatory Country of origin Text Peru countrycountryObligatoryCountry of originTextPeru Note : NA = not applicable. Note : NA = not applicable. "},{"text":"TABLE 2 . A comparison of the four index formats detailing the fi elds used in the tables format.SBMG, format.ASPT, format.NYBG, and format.PK. Systematic Botany New York Botanical Garden Systematic BotanyNew York Botanical Garden Monographs Systematic Botany publications PhytoKeys MonographsSystematic BotanypublicationsPhytoKeys Field Style Sept a Style Sept a Style Sept a Style Sept a Required b FieldStyleSept aStyleSept aStyleSept aStyleSept aRequired b species bold capitals bold bold obligatory speciesboldcapitalsboldboldobligatory country bold . none . bold;uppercase . uppercase . obligatory countrybold.none.bold;uppercase.uppercase.obligatory majorarea capitals : none : bold : none : obligatory majorareacapitals:none:bold:none:obligatory minorarea none , none , none , none , obligatory minorareanone,none,none,none,obligatory locnotes none , none , none , none , optional locnotesnone,none,none,none,optional latitude none , none , none , none , optional latitudenone,none,none,none,optional longitude none , none , none , none , optional longitudenone,none,none,none,optional altitude none , none , none , none , optional altitudenone,none,none,none,optional colldate none , none , none none , optional colldatenone,none,nonenone,optional phenology c () , optional phenology c(),optional collcite italics italics underline none obligatory collciteitalicsitalicsunderlinenoneobligatory number italics italics underline none obligatory numberitalicsitalicsunderlinenoneobligatory dups () ; () ; () ; () ; optional dups();();();();optional group.majorarea yes - no - no - no - obligatory group.majorareayes-no-no-no-obligatory species.referral ();number , ();number , ();number ; ();name ; obligatory species.referral();number,();number,();number;();name;obligatory group.specimens yes - yes - yes - yes - obligatory group.specimensyes-yes-yes-yes-obligatory "},{"text":"TABLE 3 . A comparison of the four index formats showing examples of how to create the format using the exsic function . New York Botanical Garden New York Botanical Garden Field in exsic data table Systematic Botany Monographs Systematic Botany publications PhytoKeys Field in exsic data tableSystematic Botany MonographsSystematic BotanypublicationsPhytoKeys Example exsic command a exsic(rec, format = exsic(rec, format = exsic(rec, format = exsic(rec, format = Example exsic command a exsic(rec, format =exsic(rec, format =exsic(rec, format =exsic(rec, format = format.SBMG) format.ASPT) format.NYBG) format.PK) format.SBMG)format.ASPT)format.NYBG)format.PK) Exsic formatted record Argentina. CATAMARCA : Ambato, Argentina. Catamarca: Exsic formatted recordArgentina. CATAMARCA : Ambato,Argentina. Catamarca: Sierra de Ambato, 27 ° 42 ′ S, Ambato, Sierra de Ambato, Sierra de Ambato, 27 ° 42 ′ S,Ambato, Sierra de Ambato, 65 ° 55 ′ 60 ″ W, 3500 m, 22 Feb 27 ° 42 ′ S, 65 ° 55 ′ 60 ″ W, 65 ° 55 ′ 60 ″ W, 3500 m, 22 Feb27 ° 42 ′ S, 65 ° 55 ′ 60 ″ W, 1971, Hunziker 20938 (CORD). 3500 m, 22 Feb 1971, 1971, Hunziker 20938 (CORD).3500 m, 22 Feb 1971, Hunziker 20938 (CORD). Hunziker 20938 (CORD). "}],"sieverID":"736ce53d-43a7-49c0-88cb-4a03eae5324f","abstract":""} |
