KDercksen / pmid_drug_linker

Drugs to PMIDs linker

Installation

Install Python and the required libraries:

pip install -r requirements.txt

(or use Anaconda to do this).

Usage

The tool has two main options, create and update.

create

This command is used to create a new dataset with pmid -> drug links. Open a command prompt and run the following command:

python pmid_drug_linker.py create
    --pmids literature.xlsx
    --relevant-drugs drugs.xlsx
    --output pmid_drug_mapping.csv

update

This command is used to generate a new dataset as well as a changelog given a previous dataset. Mainly useful to e.g. update a dataset with some new literature or drugs, and having the difference in a separate file.

python pmid_drug_linker.py update
    --pmids literature.xlsx
    --relevant-drugs drugs.xlsx
    --old-dataset pmid_drug_mapping.csv  # output from a previous "create" run
    --changelog-path pmid_drug_mapping_changelog.csv  # where to write difference between pmid_drug_mapping.csv and pmid_drug_mapping_new.csv
    --output pmid_drug_mapping_new.csv  # new dataset (what you would get from "create" run)

For more info on the arguments, run python find_drugs.py --help.

---

Sheet format

Example sheets are included in the resources folder, these are also used for running tests.

literature.xlsx

A	B	C	D
pmid	year	title	abstract

Examples of fields:

field	content
pmid	24936338
year	2014
title	Modeling the time dependent biodistribution of Samarium-153 ethylenediamine tetramethylene phosphonate using compartmental analysis
abstract	AIM: The main purpose of this work was to develop a pharmacokinetic model for the bone pain palliation agent Samarium-153 ethylenediamine tetramethylene phosphonate ([(153)Sm]-EDTMP) in normal rats to analyze the behavior of the complex. BACKGROUND: The use of compartmental analysis allows a mathematical separation of tissues and organs to determine the concentration of activity in each fraction of interest. Biodistribution studies are expensive and difficult to carry out in humans, but such data can be obtained easily in rodents. MATERIALS AND METHODS: We have developed a physiologically based pharmacokinetic model for scaling up activity concentration in each organ versus time. The mathematical model uses physiological parameters including organ volumes, blood flow rates, and vascular permabilities; the compartments (organs) are connected anatomically. This allows the use of scale-up techniques to predict new complex distribution in humans in each organ. RESULTS: The concentration of the radiopharmaceutical in various organs was measured at different times. The temporal behavior of biodistribution of (153)Sm-EDTMP was modeled and drawn as a function of time. CONCLUSIONS: The variation of pharmaceutical concentration in all organs is described with summation of 6-10 exponential terms and it approximates our experimental data with precision better than 2%.

The first row in this sheet should contain column headers (e.g. "PMID", "Title" ...) and will be skipped during loading.

drugs.xlsx

A	B	C	D	E
id	drug	synonym 0	synonym ...	synonym N

Examples of fields:

field	content
id	0
drug	midazolam
synonym 0	Dormicum
synonym 1	Versed
...	...

Synonyms should be alternate names for drug, one per column. The default for maximum number of synonyms is 4, you can use --num-synonyms N to use less or more. Again, the first row in the sheet should contain column headers and will be skipped during loading.

Testing

Use pytest command to run the tests (as far as there are any :-)).