Report sections 2/3 issues

[vedina]

Related to #149 The values for this report don't seem right

There should be no "medium" condition for genetic tox in vitro nor "std_dev" (it's not a condition anyway). Could you check where these come from?

The table has a column endpoint, but this should not be visible for this genetic invitro section. Also formatting should match the one as in the matrix, not a generic table as in the popup.

[gonzomir]

Can't tell why, but this data comes from the dataset. The feature has the following structure:

{
  "type":"Feature",
  "title":"",
  "units":null,
  "isNominal":false,
  "isNumeric":false,
  "isMultiValue":true,
  "sameAs":"http://www.opentox.org/echaEndpoints.owl#TO_GENETIC_IN_VITRO",
  "isModelPredictionFeature":true,
  "creator":"Just lsdfjlka sj",
  "order":1007,
  "source":{
    "URI":"https://apps.ideaconsult.net/data/dataset/Just+lsdfjlka+sj",
    "type":"read-across from supporting substance (structural analogue or surrogate)"
  },
  "annotation":[
    {"type":"conditions","p":"medium","o":"  "},
    {"type":"conditions","p":"metabolic activation","o":"  "},
    {"type":"conditions","p":"metabolic activation system","o":"  "},
    {"type":"conditions","p":"species","o":"  "},
    {"type":"conditions","p":"std_dev","o":"=3.4  mg/l"}
  ],
  "URI":"https://apps.ideaconsult.net/data/property/TOX/TO_GENETIC_IN_VITRO_SECTION//90CFAAC9D295E4A294963D1EF57B386CC01C30DC/Read-across/9a425a8d-23bc-3c9c-ba9b-a529c970081b"
}

As one can see, the annotations contain conditions "medium" and "std_dev".

Could it be a bug with adding endpoint data? Will the server accept such data?

[vedina]

Not sure why this is disturbing :)

std_dev is an artifact , but medium is ok, and annotations are fine overall . The datasets in bundles are generated from study data, and conditions and parameters go into feature annotations. Could you explain what's the issue?

[gonzomir]

The issue reported is that in annex 3 in the report the data is displayed incorrectly:

There should be no "medium" condition for genetic tox in vitro nor "std_dev" (it's not a condition anyway). Could you check where these come from?

[vedina]

could you tell which substance it is?

[gonzomir]

The substance is formaldehyde (IUC4 DSN 52) - the first one in the assessment.

[vedina]

OK, the study itself is intact, looking what's going with matrix generation.

This entry (with medium and stdev) is stored in the matrix (i.e. the bundle_substance_experiment table, which is a copy of the initial matrix, but then includes manual edits. This particular entry is a result of manual edit, this is why it does not exist in the original study. Hence, the dataset rendering is correct.

Next is to find out why the manual edit results in additional entries.

Finding:

• With the newest deployed version a manual entry doesn't generate additional conditions.
• The previous entry is likely an artifact of GENETIC_INVITRO section misconfiguration, where default parameters for each endpoint (and in early versions st_dev was stored as parameter).

To summarize:

The conditions and parameters may contain any entry, there are no fixed fields in the database. The particular entries per endpoint are configured via the study configuration. For the dataset serialisation, the conditions and parameters are written as property annotations.

Suggestion

• Leave it as it is. Removing this particular entry and inserting a new will remove the undesired conditions.
• Use the config-study.js to figure out which are the expected parameters/conditions and include only them in the report (client side) or in the dataset serialisation (server side). This has the drawback that if there are entries not listed in the configuration, the report will miss them. With IUCLID5 it's not an issue, but with any other source of data (e.g. external data or nanomaterials ) it will be a big issue. Hence my preferred suggestion is to leave it as it is.