Skip to main content

Augment's a PubMed literature search with citation data from NIH-OCC's iCite.

Project description

PubMed ID (PMID) Cite

Tweet build CodeQL Latest PyPI version DOI

pmidcite summary

Turbocharge a PubMed literature search in biomedicine, biochemistry, chemistry, behavioral science, and other life sciences by linking citation data from the National Institutes of Health (NIH) with PubMed IDs (PMIDs) using the command line rather than clicking and clicking and clicking on Google Scholar "Cited by N" links.

This open-source project is part of a peer-reviewed paper published in Research Synthesis Methods. Please cite if you use pmidcite in your research or literature search.

Contact: dvklopfenstein@protonmail.com

Usage

1) Download citation counts and data for a research paper

$ icite -H 26032263

  • This paper (PMID 26032263) has 25 citations, 10 references, and 4 authors.
  • This paper is performing well (74th percentile in column %) compared to its peers.

Starting usage

NIH percentile

This paper is performing well (74th percentile) compared to its peers (column %).

The NIH percentile grouping (column G) helps to highlight the better performing papers in groups 2, 3, and 4 by sorting the citing papers by group first, then publication year.

The sort places the lower performing papers in groups 0 or 1 at the back.

New papers appear at the beginning of a sorted list, no matter how many citations they have to better facilitate researchers in finding the latest discoveries.

The grouping of papers by NIH percentile grouping is a novel feature created by dvklopfenstein for this project.

2) Forward citation search

Also known as following a paper's Cited by links or Forward snowballing

icite -H; icite 26032263 --load_citations | sort -k6 -r
or
icite -H; icite 26032263 -c | sort -k6 -r

3) Backward citation search

Also known as following links to a paper's references or Backward snowballing

$ icite -H; icite 26032263 --load_references | sort -k6 -r
or
$ icite -H; icite 26032263 -r | sort -k6 -r

PubMed vs Google Scholar

Google Scholar vs PubMed

In 2013, Boeker et al. recommended that a scientific search interface contain five integrated search criteria. PubMed implements all five, while Google did not in 2013 or today.

Google's highly popular implementation of the forward citation search through their ubiquitous "Cited by N" links is a "Better" experience than the PubMed's "forward citation search" implementation.

But if your research is in the health sciences and you are amenable to working from the command line, you can use PubMed in your browser plus citation data downloaded from the NIH using the command-line using pmidcite. The NIH's citation data includes a paper's ranking among its co-citation network.

What is in PubMed? Take a quick tour

PubMed Contents

PubMed is a search interface and toolset used to access over 30.5 million article records from databases such as:

  • MEDLINE: a highly selective database started in the 1960s
  • PubMed Central (PMC): an open-access database for full-text papers that are free of cost
  • Additional content such as books and articles published before the 1960s

Usage details

Download citations for all papers returned from a PubMed search

Make a copy of src/bin/dnld_pmids.py and add your PubMed search to the end of the queries list.

There are two PubMed searches in this example:

  • systematic review AND "how to"[TI]
  • Orcinus Orca Type D

The PubMed search results are saved to specified filenames such as systematic_review.txt to be grepped and sorted.

def main():
    """Download PMIDs returned for a PubMed query. Write an iCite report for each PMID"""
    queries = [
        # Output filenames               PubMed query
        # -----------------              -----------------------------------
        ('systematic_review.txt',        'systematic review AND "how to"[TI]'),
        ('rarely_seen_killer_whale.txt', 'Orcinus Orca Type D'),
    ]

    obj = PubMedQueryToICite(force_dnld=True)
    dnld_idx = obj.get_index(sys.argv)
    obj.run(queries, dnld_idx)

To have better access to PubMed search results, get a NCBI API key using these instuctions:
https://ncbiinsights.ncbi.nlm.nih.gov/2017/11/02/new-api-keys-for-the-e-utilities

Table of Contents

Command Line Interface (CLI)

A Command-Line Interface (CLI) can be preferable to a Graphical User Interface (GUI) because:

  • processing can be automated from a script
  • time-consuming mouse clicking is reduced
  • more data can be seen at once on a text screen than in a browser, giving the researcher a better overall impression of the full set of information [1]

Researchers who use Linux or Mac already work from the command line. Researchers who use Windows can get that Linux-like command line feeling while still running native Windows programs by downloading Cygwin from https://www.cygwin.com/ [1].

1) Get citation counts, given PMIDs

Quickly get the number of citations for a research paper with PMID, 26032263:

$ icite 26032263 -H
TYP PMID     RP HAMCc   % G YEAR   cit cli ref au[00](authors) title
TOP 26032263 R. .....  68 2 2015    16  0  10 au[04](N R Haddaway) Making literature reviews more reliable through application of lessons from systematic reviews.
  • The first line (TYP PMID ...) contains the column headers (-H).
  • The second line (TOP ...) is the citation data from NIH's iCite database.
  • The citation count, 16, is under the cit column header.

The group number, 2 (SD column) indicates that the paper has a good citation rate, specifically it is in the 68th percentile (% column) compared to its peers.

Column header key (-k)

$ icite -k

KEYS TO PAPER LINE:
    TYP PubMedID RP HAMCc % G YEAR x y z au[A](First Author) Title of paper

TYPe of relationship to the researcher-requested paper (TYP):
    TOP: The paper requested by the researcher
    CIT: A paper that cited TOP
    CLI: A clinical paper that cited TOP
    REF: A paper referenced in the TOP paper's bibliography

NIH iCite details:

  PubMedID: PubMed ID (PMID)

     RP section:
     ----------------------------------
         R: Is a research article
         P: iCite has calculated an initial Relative Citation Ratio (RCR) for new papers

     HAMCc section:
     ----------------------------------
         H: Has MeSH terms in the human category
         A: Has MeSH terms in the animal category
         M: Has MeSH terms in the molecular/cellular biology category
         C: Is a clinical trial, study, or guideline
         c: Is cited by a clinical trial, study, or guideline

     NIH section, based on Relative Citation Ratio (RCR):
     ----------------------------------
         %: NIH citation percentile rounded to an integer. -1 means "not determined" or TBD
         G: NIH citation percentile group: 0=-3SD 1=-2SD 2=+/-1SD 3=+2SD 4=+3SD or i=TBD

     YEAR/citations/references section:
     ----------------------------------
      YEAR: The year the article was published
         x: Total of all unique articles that have cited the paper, including clinical articles
         y: Number of unique clinical articles that have cited the paper
         z: Number of references
     au[A]: A is the number of authors

Citation group numbers [1]

The pmidcite citation rate group numbers, 0, 1, 2, 3, and 4 (SD column), are determined using the NIH Relative Citation Rate (RCR) [5] percentile. If the NIH has not yet determined a citation rate for new papers, the pmidcite group number is i.

cite group

2) Sort citation counts, given PMIDs

Sort the citations (CIT) of the paper with PMID 26032263 first by citation group (2 and i), then by year.

The citation group shown contains:

  • i New paper and not yet rated. The i variable will be set at a later date by the NIH
  • 2 These papers are performing well

Sort options:

  • -k6: sort starting with the 6th column containing citation group, then by all text to the right.
  • -r: reverse the sort so the newest papers are at the top
$ icite 26032263 -v | grep CIT | sort -k6 -r
CIT 32557171 .. H....  -1 i 2020     0  0  21 au[05](Jillian Knox) Usage, definition, and measurement of coexistence, tolerance and acceptance in wildlife conservation research in Africa.
CIT 32317639 R. HA...  -1 i 2020     0  0   8 au[09](Trevor J Krabbenhoft) FiCli, the Fish and Climate Change Database, informs climate adaptation and management for freshwater fishes.
CIT 30285277 R. .....  -1 i 2019     2  0  14 au[02](Neal R Haddaway) Predicting the time needed for environmental systematic reviews and systematic maps.
CIT 30055022 .. HA...  -1 i 2019     1  0  12 au[04](Hillary Smith) Hunting for common ground between wildlife governance and commons scholarship.
CIT 31598307 R. HA...  -1 i 2019     1  0  12 au[02](Igor Khorozyan) How long do anti-predator interventions remain effective? Patterns, thresholds and uncertainty.
CIT 31024221 R. .....  -1 i 2019     0  0   7 au[02](Micah G Bennett) MEASURING LOTIC ECOSYSTEM RESPONSES TO NUTRIENTS: A Mismatch that Limits the Synthesis and Application of Experimental Studies to Management.
CIT 29488217 .P .A...  76 2 2018     7  0  64 au[03](Nicole V Coggan) A global database and 'state of the field' review of research into ecosystem engineering by land animals.
CIT 29514874 .P .A...  47 2 2018     3  0  38 au[02](Kelly D Hannan) Aquatic acidification: a mechanism underpinning maintained oxygen transport and performance in fish experiencing elevated carbon dioxide conditions.
CIT 28642071 .. H....  75 2 2017    11  0  80 au[05](Ora Oudgenoeg-Paz) The link between motor and cognitive development in children born preterm and/or with low birth weight: A review of current evidence.
CIT 28061344 R. .....  70 2 2017     8  0  54 au[03](Maria Cristina Mangano) Monitoring of persistent organic pollutants in the polar regions: knowledge gaps & gluts through evidence mapping.
CIT 28042667 R. H....  53 2 2017     8  0  20 au[02](Martin J Westgate) The difficulties of systematic reviews.
CIT 29451529 .. H....  56 2 2016     9  0  20 au[01](Jennifer A Byrne) Improving the peer review of narrative literature reviews.
CIT 26984257 R. .....  46 2 2016     9  0   9 au[04](Neal R Haddaway) The benefits of systematic mapping to evidence-based environmental management.
CIT 27617203 .. .....  43 2 2016     5  0  40 au[02](Neal R Haddaway) On the benefits of systematic reviews for wildlife parasitology.

Other sort examples

In 2018 Fiorini et al. [7], the creaters of PubMed's "best match" relevance sort ordering in PubMed, found that the most important document features to feed into the PubMed sorting algorithm are publication year and past usage.

Mimic this by using the -k6 argument to sort the citation group (usage group), which does two things:

  • First, it highlights the newest or best performing papers by putting them at the beginning, while getting the lowest performing papers out of the mix by placing them at the end.
  • Second, it shows the newest papers first in each usage group, highlighting them profoundly.

We chose to highlight using usage group first, rather than NIH RCR percentile in the 5th column, seen with values -1, 76, etc. because only seeing the best performing papers first might bias the paper chosen for further examination to only the best performing papers regardless of publication year.

3) Query PubMed and download the citation data

Query PubMed and download the citation data from the script, src/bin/dnld_pmids.py.
NOTE: Copy dnld_pmids.py to your project repo. Don't modify the pmidcite repo.

1. Add your query to your dnld_pmids.py script

    queries = [
        # Output filename     PubMed query
        # -----------------  -----------------------------------
        ('killer_whale.txt', 'Orcinus Orca Type D'),
    ]

2. Run the script

$ src/bin/dnld_pmids.py
     3 IDs FOR pubmed QUERY(Orcinus Orca Type D)
     3 WROTE: ./log/pmids/killer_whale.txt
     3 WROTE: ./log/icite/killer_whale.txt

3. Examine the citation and pubmed data, sorting by year (column 7; -k7)

$ grep TOP ./log/icite/Orcinus_Orca_Type_D.txt | sort -k7
TOP 20050301 R. .A...  70 2 2009    43  0  25 au[05](Andrew D Foote) Ecological, morphological and genetic divergence of sympatric North Atlantic killer whale populations.
TOP 22882545 .. .A...  63 2 2013    25  0  24 au[03](P J N de Bruyn) Killer whale ecotypes: is there a global model?
TOP 31461780 R. .A...  -1 i 2020     0  0   0 au[06](Robert L Pitman) Enigmatic megafauna: type D killer whale in the Southern Ocean.

4) Get citation data using PMIDs downloaded from PubMed

Note that the PubMed query using NIH E-Utils from the dnld_pmids.py script will often be slightly different than the query run on the PubMed website. PubMed has been alerted.

Consequently, you may also want to view citation data on PMID PubMed query results downloaded from the PubMed website into a file such as pmid-OrcinusOrc-set.txt:
Save->All results, Format=PMID

$ icite -i pmid-OrcinusOrc-set.txt
TOP 30123694 RP HA...  17 2 2018     1  0   6 au[07](Paul Tixier) Killer whale (<i>Orcinus orca</i>) interactions with blue-eye trevalla (<i>Hyperoglyphe antarctica</i>) longline fisheries.
TOP 31461780 R. .A...  -1 i 2020     0  0   0 au[06](Robert L Pitman) Enigmatic megafauna: type D killer whale in the Southern Ocean.
TOP 22882545 .. .A...  63 2 2013    25  0  24 au[03](P J N de Bruyn) Killer whale ecotypes: is there a global model?
TOP 20050301 R. .A...  70 2 2009    43  0  25 au[05](Andrew D Foote) Ecological, morphological and genetic divergence of sympatric North Atlantic killer whale populations.

5) Create ASCII plots

Create a scatter plot of publication year vs. citation count for a list of papers. This will be made friendlier.

Columns 7 and 8 contain the year and the citation count.

$ grep TOP log/icite/Osbourn_Anne.txt | awk '{print $7 " " $8}' | scatter.py
-------------------------------------------------------------------------------------------- 282
|                                                                1                         |
|                                                                                          |
|                                                                                          |
|                                                                                          |
|                                                                                          |
|                                                                                          |
|                                                                                          |
|                                                                                          |
|                              1                                                           |
|                                                                                          |
|               1                                                                          |
|                                                                                          |
|                                                                                          |
|                                                           1                              |
|          1                                                     1                         |
|1                                  1                                                      |
|                                                                                          |
|                                                      1                                   |
|                    1                   1                                                 |
|                                            1                   1                         |
|                                            1              1    1                         |
|     1                             1                                 1                    |
|                    1              1                  1    1                              |
|                              1             1              1                              |
|     1              1              1    1        2    2                                   |
|                                        1                            1    2               |
|                                                           1    2    1    1         1     |
|1    1                                  1   2    1    1              3    1    4          |
|               1         1    1         1   1    3    1    1    1              5          |
|          2         2                                           1    1    2    1    7    3|
-------------------------------------------------------------------------------------------- 0
2002                                                                                          2020

Installation

To install from PyPI
pip install pmidcite

Setup

Save your literature search in a GitHub repo.

1. Add a pmidcite init file

Add a .pmidciterc init file to a non-git managed directory, such as home (~)

$ icite --generate-rcfile | tee ~/.pmidciterc
[pmidcite]
email = name@email.edu
apikey = long_hex_digit
tool = scripts
dir_icite_py = .
dir_pubmed_txt = .
dir_pmids = .
dir_icite = .
$ export PMIDCITECONF=~/.pmidciterc

Do not version manage the .pmidciterc using a tool such as GitHub because it contains your personal email and your private NCBI API key.

2. Add directories

Add directories which match those in ~/.pmidciterc:

$ mkdir [GIT_REPO_PATH]/icite
$ mkdir [GIT_REPO_PATH]/log
$ mkdir [GIT_REPO_PATH]/log/pubmed
$ mkdir [GIT_REPO_PATH]/log/pmids
$ mkdir [GIT_REPO_PATH]/log/icite

3. NCBI E-Utils API key

To download PubMed abstracts and PubMed search results using NCBI's E-Utils, get an NCBI API key using these instructions:
https://ncbiinsights.ncbi.nlm.nih.gov/2017/11/02/new-api-keys-for-the-e-utilities

Set the apikey value in the config file: ~/.pmidciterc

Contact

email: dvklopfenstein@protonmail.com
https://orcid.org/0000-0003-0161-7603

How to Cite

If you use pmidcite in your research or literature search, please cite paper 1 (pmidcite) and paper 3 (NIH citation data).

Please also consider reading and citing Gusenbauer's response (paper 2) about improving search for all during the information avalanche of these times:

  1. Commentary to Gusenbauer and Haddaway 2020: Evaluating Retrieval Qualities of PubMed and Google Scholar
    Klopfenstein DV and Dampier W
    2020 | Research Synthesis Methods | PMID: 33031632 | DOI: 10.1002/jrsm.1456 | pdf

  2. Gusenbauer's response:
    What every Researcher should know about Searching – Clarified Concepts, Search Advice, and an Agenda to improve Finding in Academia
    Gusenbauer M and Haddaway N
    2020 | Research Synthesis Methods | PMID: 33031639 | DOI: 10.1002/jrsm.1457 | pdf

  3. The NIH citation data used by pmidcite -- Scientific Influence, Translation, and Citation counts:
    The NIH Open Citation Collection: A public access, broad coverage resource
    Hutchins BI ... Santangelo GM
    2019 | PLoS Biology | PMID: 31600197 | DOI: 10.1371/journal.pbio.3000385

References

Please consider reading and citing the paper [4] which inspired the creation of pmidcite [1] and the authors' response to our paper [2]:

  1. Which Academic Search Systems are Suitable for Systematic Reviews or Meta-Analyses? Evaluating Retrieval Qualities of Google Scholar, PubMed and 26 other Resources
    Gusenbauer M and Haddaway N
    2019 | Research Synthesis Methods | PMID: 31614060 | DOI: 10.1002/jrsm.1378

Mentioned in this README are also these outstanding contributions:

  1. Relative Citation Ratio (RCR): A New Metric That Uses Citation Rates to Measure Influence at the Article Level
    Hutchins BI, Xin Yuan, Anderson JM, and Santangelo, George M.
    2016 | PLoS Biology | PMID: 27599104 | DOI: 10.1371/journal.pbio.1002541

  2. Google Scholar as replacement for systematic literature searches: good relative recall and precision are not enough
    Boeker M et al.
    2013 | BMC Medical Research Methodology | PMID: 24160679 | DOI: 10.1186/1471-2288-13-131

  3. Best Match: New relevance search for PubMed
    Fiorini N ... Lu Zhiyong
    2018 | PLoS Biology | PMID: 30153250 | DOI: 10.1371/journal.pbio.2005343

PDFs

Contact

dvklopfenstein@protonmail.com
https://orcid.org/0000-0003-0161-7603

Copyright (C) 2019-present pmidcite, DV Klopfenstein. All rights reserved.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

pmidcite-0.0.36.tar.gz (2.6 MB view hashes)

Uploaded source

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page