Phenodigm results

We present some examples on how to construct a URL to retrieve features of interest and then download from that URL.

Here we look for phenodigm scores for Sod1 (mouse gene identifier MGI:98351):

https://www.ebi.ac.uk/mi/impc/solr/phenodigm/select?rows=250000&q=type:disease_model_summary%20AND%20marker_id:”MGI:98351″

When pasting the above URL in a browser (all on one line, watch out for the quotation marks), the browser will retrieve data for that gene. The output comes in json format by default.

You can request to get the output in table format instead. It is also possible to fetch specific columns.

https://www.ebi.ac.uk/mi/impc/solr/phenodigm/select?rows=250000&q=type:disease_model_summary%20AND%20marker_id:”MGI:98351″&wt=csv&fl=marker_id,model_id,model_description,model_genetic_background,disease_id,disease_term,association_curated,disease_model_avg_norm,disease_model_max_norm

The search can be restricted to specific diseases:

https://www.ebi.ac.uk/mi/impc/solr/phenodigm/select?rows=250000&q=type:disease_model_summary%20AND%20marker_id:”MGI:98351″%20AND%20disease_id:”OMIM:105400″&wt=csv&fl=marker_id,model_id,model_description,model_genetic_background,disease_id,disease_term,association_curated,disease_model_avg_norm,disease_model_max_norm

The output has columns “disease_model_avg_norm” and “disease_model_max_norm”. Please note that all the scores describe associations between diseases and animal models, not genes. (We define an animal model as a combination of strain, allele type, and zygosity; for IMPC models, also life stage.)

To download data for more genes and diseases, just replace the identifiers in the URLs. Alternatively, you may prefer to download all the data at once:

https://www.ebi.ac.uk/mi/impc/solr/phenodigm/select?rows=1&q=type:disease_model_summary&wt=csv&fl=marker_id,model_id,model_description,model_genetic_background,disease_id,disease_term,association_curated,disease_model_avg_norm,disease_model_max_norm

The above is rigged with “rows=1” to display a single row. To really download the whole dataset, the row count needs to be replaced with a number larger than 10 million. A utility like curl can download the ~1GB table directly into a file.