litspy 0.0.5

Searches through all titles and abstracts available in Europe PMC for co-occurrence of supplied terms (and their synonyms where available), and produces summaries of the search results.

LitSpy

DESCRIPTION

This package allows the user to search through all titles and abstracts available in Europe PMC for articles containing co-occurence of the following:

• supplied genes and their synonyms
• an optionally supplied disease and its synonyms
• an optionally supplied tissue or organ and its synonyms
• optionally other supplied keywords
• and any other Europe PMC advanced search parameters desired

and produces summaries of the search results in HTML and optionally Excel format.

INSTALLATION

The tool is available from pypi, and can be easily installed and upgraded in the usual way using pip. Just ensure that you have Python3.6 or higher installed on your machine, and run the command

pip install litspy

(or, if your default Python version is Python2, replace pip with pip3)

USAGE

Getting started

Note: replace python with python3 in all commands if your default python version is not python3

Obtaining usage information

To read about the different arguments you can supply to LitSpy (such as input files, output files, logging settings and more), run

python -m litspy -h

Basic commands (examples)

Unless specified otherwise, LitSpy assumes a UniProt gene ID for a human gene has been entered. Therefore, a minimal search can simply include a human gene ID and a disease, tissue and/or keyword. For example:

python -m litspy -g Cftr -d arthritis

The command can include multiple genes and multiple keywords (separated by spaces), but only one disease and one tissue/organ. To search for phrases that include spaces, you should use quotes around the phrases:

python -m litspy -g APOE APOC1 APOC2 -d "cystic fibrosis" -t liver -k "normal diet" biomarker

To supply a Uniprot accession number instead of a gene ID, specify the UniProt ID type as accession using -u or --uniprot_id_type, for example:

python -m litspy -g P13569 -u accession -d arthritis

To search for non-human genes, specify the appropriate UniProt taxonomy ID using s or --taxid, for example:

python -m litspy -g CFTR -d arthritis -s 10090

Using input files

To supply an input file, use -i (or --infile) followed by the path to the input file, for example:

python -m litspy -i /home/myusername/documents/litspy_inputs/input1.xlsx

LitSpy will preferentially take an input file over values supplied at the command line. If the following command was entered, only the input file would be used and an appropriate warning would be logged

python -m litspy -i path/to/input.xlsx -g CFTR APOE

Input files should be spreadsheets in xlsx format, containing the following columns in the following order:

UniProtID	IDType	TaxID	Keyword
CFTR	gene	9606	normal diet, biomarker
CFTR	gene	9606	children
APOE	gene	9606	normal diet, biomarker
P13569	accession	9606	biomarker

Additional columns to the right are tolerated, but ignored. (They may be used for notes, for example)

LitSpy runs one query per row in the spreadsheet (ignoring the headings).

To query for a gene co-occurring with multiple keywords, enter the keywords as a comma-separated list in the 4th column of the input file (note: quote marks are not needed in input files). To query for a gene that co-occurs with any of a list of keywords, you should enter separate rows for each keyword. So, for example, the input file above would return documents containing CFTR AND normal diet AND biomarker, and also documents containing CFTR AND children.

Create an Excel output file

To get an output file of the results summary and details in Excel format, use -o (or --outfile) followed by the full path to an output file:

python -m litspy -g CFTR -d arthritis -o /home/myusername/documents/litspy_outputs/output1.xlsx

If you provide a path to an existing file, the file will be overwritten. If you provide a path to a new file, the new file will be created.

The Excel output file is created in addition to the HTML results files, which are always created.

Create output charts and top 10 list

Output charts are generated automatically if fewer than 30 genes are supplied. If more than 30 genes are supplied, charting is turned off to prevent potential creation of 90+ images. To turn charting on or off despite input size, use the -c or --make-charts flag, with 'y' or 'n' to turn charting on or off respectively.

A list of the ten most common terms (excluding search terms) can be generated for each result set by specifying the -w or --top-ten flag with 'y'.

For example:

python -m litspy -i 'path/to/input.xlsx -w y -c y

will create result pages with charts and top ten lists no matter the number of entries in the supplied file

Logging

By default, LitSpy logs to the console only at the warning level. To change the console logging level, supply a log level argument by entering -l (or --log-level) followed by a valid logging level, for example:

python -m litspy -g CFTR -d arthritis -l info

To turn on logging to a log file, specify the -f (or --log-file) flag:

python -m litspy -g CFTR -d arthritis -l warning -f

The log file is always logged at the info level.

Multithreading

Some parts of the tool are able to run in parallel on many threads. You can specify the number of threads to use with the -m or --multithread flag, for example:

python -m litspy -g CFTR -d arthritis -m 4

If you do not specify the number of threads to use, then the tool will automatically determine the number of available cores and use this as the maximum number of threads to use. It is possible that this will affect performance of other tasks that the machine is running at the same time.

Keyword synonym expansion

To attempt collection of synonyms for supplied keywords, specify the -e (or --expand-keywords) flag in the command:

python -m litspy -g CFTR -d arthritis -k biomarker -e

This will return the synonyms from the first node matched (the best match) when searching for the term in the EBI OLS.

Warning: It is expected that this generic searching will be far noisier on average than the more specific searches for disease or tissue synonyms, and therefore this option should be used cautiously.

BACKGROUND

The aim of this tool is to form part of my PhD project, where it will help to inform target novelty within particular contexts, and be combined with other tools. However, it can also be useful as a standalone search tool; for example, to identify articles of interest more comprehensively than by keyword searches that do not use synonym expansion.

AUTHOR

Emma Croot, 2020, [e-mail], github.com/ec339

LICENCE

MIT (see LICENCE.txt)

CHANGE LOG

There have currently been no major changes to the tool. See commits at github for information about minor updates.

ACKNOWLEDGEMENTS

OmixLitMiner

Although the design and creation of this tool began before Pascal et al's 2020 publication (DOI: 10.3390/ijms21041374, PMID: 32092871), their tool OmixLitMiner provided inspiration for parts of this project; particularly the use of UniProt to obtain alternative official gene names, and the output charts and styles.

Supervision team

My PhD supervisors, Dr. Thanos Didangelos, Dr. Richard Badge, and Prof. Louise Wain

Others

Thomas Rowlands, co-developer of rtgo (used in LitSpy for multithreading)