IsoMemo: a Big isotopic Data initiative

Our goal is to serve isotope data producers and users from multiple research disciplines with vastly different isotope tracers, research questions, measurement techniques, sample types, environmental settings, time period of interest, and availability of complementary sources of data, etc. In view of this, the approach of IsoMemo is to define several data tables instead of a single universal table to accommodate all different perspectives. The definition of the different tables is made according to the needs of the users within the different isotope data communities.  During a first discussion round users can provide feedback (suggest new tables, merger of tables, organization of sub-tables, fields, etc.) through the IsoMemo discussion group. The following lists a small number of tables already open for discussion:

“AMS 13C” table containing 13C data measured using AMS systems

“Ancient bone” table of isotope data obtained from ancient bone (contains example data)

“Ancient ceramics” table of isotope data obtained from ancient ceramics (contains example data)

“Ancient plant remains” table of isotope data obtained from ancient plant remains

“Ancient teeth” table of isotope data obtained from ancient teeth

“Feeding experiments” table of isotope data from feeding experiments

“Modern human hair and nails” table of isotope data from modern human hair and nails (contains example data)

A separate table stores user’s data: “Users”

Other tables are currently being designed. These include, modern plants; modern plant leaf wax; Sr isotopie data (measured on modern water, rock, and soil samples), compound specific amino acid of human and other organisms, and others.

The tables can be downloaded as Excel files that contain a list of fields, metadata, and in a few cases already some data. The fields are organized into thematic sub-tables. These are not real tables but will facilitate the process of navigating and querying the database once the advanced user interface is implemented. The tables can also be used to submit data by providing a data-populated field to and following the instructions provided within the files.

Isotopic measurements represent the combination of the “isotopes” of choice measured on a certain sample “sample type”. For example, 13C/12C can be measured in bulk bone collagen or bioapatite. However, single amino acids isolated from bone collagen can also be subjected to 13C/12C measurements and this applies also to other isotopic proxies. Thus, depending on the particular table there could be multiple combinations of “isotopes” x ”sample type”. Thus, different approaches are open for discussion:

1) All possible combinations of “isotopes” x ”sample type” are defined for a particular table. This approach is partly illustrated in the temporary version of the “Ancient bone” table. The main disadvantage of this approach is that the number of columns can become very large. The “Ancient bone” table contains 92 fields (columns) and these do not yet include the compound specific fields.

2) “Sample type” and “isotopes” are represented in separate fields. This reduces the necessary number of fields but implies that for a single target (e.g. consumer) the data may have to be listed in several rows. The table “Ancient teeth” illustrates this approach. The tables “Ancient bone” and “Ancient teeth” are actually quite similar but the “Ancient teeth” table contains several fields that allow the user to pinpoint on what the measurement was made. This includes the field “Fraction” where the user identifies the biochemical fraction (e.g. enamel phosphate or carbonate, bulk dentin, single amino acids).

3) In some cases both the number of columns and rows for a single target can become very large. In feeding experiments it is necessary to record both the isotopic signals of food and consumer. A possible approach to this is illustrated in the table “Feeding experiments”. The isotopic signals measured in the consumers are represented in a similar way to that described for the table “Ancient ceramics”. However, nutrient data isotopic is provided in “Nutrient X signals” following a certain convention to represent the multiple measurement possibilities (e.g. 13C/12C bulk = -20.0‰, 15N/14N bulk = 6‰, 13C/12C lipids = -27.0‰, 13C/12C alanine = -26‰, 13C/12C 18:0 = -30‰). The main disadvantages of this approach are that all users have to adhere to a strict data formatting convention and following the download of the file it may require additional text operations to extract the data and have it in a tabular form.