Citation correction methods

As previously described, the PubMed Central Open Access subset of journal articles yielded 6,529,815 independent bibliographic records of both citing and cited entities, while our use of the PubMed Entrez API provided a further 2,304,143 bibliographic records for the same cited entities. Before converting these references into RDF to create the Open Citations Corpust, we attempted to remove errors in the data.

Some of the references we collected were to highly cited papers, while 2,505,879 referenced papers were only cited once. Figure 1 shows the number of citations per paper for the 100 most highly cited papers in our records – the left hand end of what is a classical long-tail dataset.

 Figure 1

We have not yet analysed the topics of these papers, but can reveal that the paper most highly cited from within the OASS, with 2150 citations, is

Altschul et al. (1997). Gapped BLAST and PSI-BLAST: a new generation of protein database. Nucleic Acids Res. 25(17):3389-3402. doi: 10.1093/nar/25.17.3389.

In an ideal world, all OASS references to an individual paper would be identical, and would exactly match the data on that paper extracted from the Entrez API. However, as we have already seen for author names, this is not the case. As with most datasets in the world, a significant proportion (~1%) of our input reference data is either incomplete or erroneous.

We attempted to correct for these errors by comparing references that appeared to reference the same bibliographic entity, and from this comparison extract the correct data for authors, title, etc. using the following rules:

  1. Accept the longest author list and names bearing accents over those lacking them.
  2. Accept DOIs and PubMed IDs from references having them, after eliminate mis-formed identifiers (e.g. DOIs lacking the journal prefix “10.****”), using the majority vote if different identifiers were given for the same paper.
  3. Accepting those variants of titles, journal names, etc. held in common by the majority of references.

This voting method was weighted in favour of data we judged to be most reliable, namely the PubMed records returned from the Entrez API, and metadata about cited papers that were within the OASS coming from the independent bibliographic records we had for those papers.

As a result of these activities, not only did we coalesce the independent references from different OASS articles to the same multiply cited papers into a set of 3,578,598 unique bibliographic citation target records describing 204,637 OASS articles and 3,373,961 articles outside the OASS, but we were able to select from the multiple references those elements (author list, title, etc.) that were judged to be correct for each target.

However, of the 2,505,879 papers that are only cited once from within the OASS, 1,246,967 lacked a PubMed ID, so for these we were unable to gather confirmatory evidence for the accuracy of the citation from the Entrez API. These references, which are to the least significant papers in the corpus, are therefore provided “as is” from PubMed Central, without any external corroboration of their accuracy.

How these error-correction processes fitted into the data processing pipeline used to create the Open Citations Corpus is described in the next blog post.

Who wrote this paper? Author list problems in PubMed Central references

To illustrate three kinds of problems in obtaining correct author lists for Open Citation data from articles in the PubMed Central Open Access subset (OASS), I take three examples, the first of which is the result of a publication policy, the second due to mis-handling of an authorship attribution at the time of publication, and the third exemplifing errors introduced when handling non-English personal names.

Example 1

In the paper by Reis et al. (2008) [1], that we took as the subject for our exercise in semantic publishing enhancements described in [2], we find the following entry for Reference 40 in the reference list:

40.    Maciel EAP, Carvalho ALF, Nascimento SF, Matos RB, Gouveia EL, et al. (2008) Household transmission of Leptospira infection in urban slum communities. PLoS Negl Trop Dis 2: e154. doi:10.1371/journal.pntd.0000154.

Note that it is the policy of the publisher, the Public Library of Science, to list only the first five authors in references to papers that have more than five, despite publishing online-only journals where article length is not an issue.

The XML for this reference in the document is as follows:

    <meta name="citation_reference" content="citation_title=Household transmission of Leptospira infection in urban slum communities.; citation_author=EAP Maciel; citation_author=ALF Carvalho; citation_author=SF Nascimento; citation_author=RB Matos; citation_author=EL Gouveia; citation_journal_title=PLoS Negl Trop Dis; citation_volume=2; citation_number=40; citation_pages=e154. doi:10.1371/journal.pntd.0000154; citation_date=2008; " /> 

Note that in the XML, all indication that there are more than five authors, i.e. the “et al.” present in the human-readable reference, is totally lost. There is thus no way of telling from the XML for this paper retrieved from PubMed Central that the full authorship for this paper is as follows:

Elves A. P. Maciel, Ana Luiza F. de Carvalho, Simone F. Nascimento, Rosan B. de Matos, Edilane L. Gouveia, Mitermayer G. Reis and Albert I. Ko.

The last two authors of the cited paper, Mitermayer Reis and Albert Ko, who are the lead author and the senior author, respectively, of the citing paper, are both omitted from the PLoS reference to the cited paper, and hence from the data automatically extracted by Open Citations from the OASS.

Example 2

A second example from the same citing paper is the cited reference to Ko et al. (1999) [2]. As the reference at the foot of this page shows, the author list includes six names “and the Salvador Leptospirosis Study Group”. Group attributions of this kind are commonplace, particularly in papers resulting from large collaborative projects. However, conventional markup systems such as the NLM-DTD have no systematic way of handling such information. Surprisingly, it is even incorrectly stated in the human readable version of the reference in the Reis et al. paper. Reference 6 in the article’s reference list reads:

6.    Ko AI, Reis MG, Ribeiro Dourado CM, Johnson WD Jr, Riley LW (1999) Urban epidemic of severe leptospirosis in Brazil. Salvador Leptospirosis Study Group. Lancet 354: 820–825.

thus including “Salvador Leptospirosis Study Group” as part of the title, an error also present in the XML version of the paper:

<meta name="citation_reference" content="citation_title=Urban epidemic of severe leptospirosis in Brazil. Salvador Leptospirosis Study Group.; citation_author=AI Ko; citation_author=MG Reis; citation_author=CM Ribeiro Dourado; citation_author=WD Johnson; citation_author=LW Riley; citation_journal_title=Lancet; citation_volume=354; citation_number=6; citation_pages=820-825; citation_date=1999; " /> 

Example 3

The third example, also taken from the reference list of Reis et al. (2008) [1], illustrated the problems of handling non-English names and titles. Reference 39 in the PLOS article’s reference list [3] reads:

Dias JP, Teixeira MG, Costa MC, Mendes CM, Guimaraes P, et al. (2007) Factors associated with Leptospira sp infection in a large urban center in Northeastern Brazil. Rev Soc Bras Med Trop 40: 499–504.

Note that the author list is truncated, that the generic name “Leptospira sp” in the title is italicized, and that neither the DOI nor the PubMed ID is provided, although the article has both.

The landing page for this article on the publisher’s web site ( shows the following:

Revista da Sociedade Brasileira de Medicina Tropical

Print version ISSN 0037-8682

Rev. Soc. Bras. Med. Trop. vol.40 no.5 Uberaba Oct. 2007

doi: 10.1590/S0037-86822007000500002  ARTIGO ARTICLE

Factors associated with Leptospira sp infection in a large urban center in northeastern Brazil

Fatores associados à infecção por Leptospira sp em um grande centro urbano do Nordeste do Brasil

Juarez Pereira DiasI; Maria Glória TeixeiraI; Maria Conceição Nascimento CostaI; Carlos Maurício Cardeal MendesI; Patrícia GuimarãesII; Mitermayer Galvão ReisII; Albert KoII,III; Maurício Lima BarretoI

The paper is published in English. Note that there is an alternative Portuguese title, that the generic name “Leptospira sp” is italicized in both, and that a DOI is provided, although page numbers are not give for this on-line version of the article. Note also the accents and structures of the full Brazilian author names.

Clicking on the tab Article in PDF format on the landing page takes one to the PDF download page that gives the following reference with page numbers, but lacking English title and DOI, and lacking the full author list:

DIAS, Juarez Pereira et al. Factors associated with Leptospira sp infection in a large urban center in northeastern Brazil. Rev. Soc. Bras. Med. Trop. [online]. 2007, vol.40, n.5, pp. 499-504. ISSN 0037-8682.

A manual search for the same article in PubMed returns the following information:

Rev Soc Bras Med Trop. 2007 Sep-Oct;40(5):499-504.

Factors associated with Leptospira sp infection in a large urban center in northeastern Brazil.

Dias JP, Teixeira MG, Costa MC, Mendes CM, Guimarães P, Reis MG, Ko A, Barreto ML.

PMID: 17992402

Note the correct accentuation of the surname “Guimarães”, but the loss of the last Christian Name initial for Maria Conceição Nascimento Costa, and the loss of italicization of the generic name “Leptospira sp“in the tile. Note also the absence of the Portuguese title and the DOI, and the addition of a PubMed ID.

The real problems with this reference arise when we look at the starting XML corpus upon which our linked open citation data output is based, which in turn is based on the original PLoS submission of the Reis et al. (2008) paper to PubMed Central.

The author list for this reference #39 in the PMC XML for Reis et al (2008) is as follows:

<ref id=“pntd.0000228-Dias1”>
<citation citation-type=“journal”>
<person-group person-group-type=“author”>

Here we see, as in the HTML reference list of the original Reis et al. (2008) paper, the truncation of the list of authors to the first five, and the loss of accent on the surname “Guimarães”. Surprisingly, the title is recorded as

<article-title>Factors associated with <italic>Leptospira</italic> sp infection in a large urban center in Northeastern Brazil.</article-title>

correctly showing the italic “Leptospira” but also correctly not italicizing the following “sp”!

All is not lost, however, in terms of the full author list. Since the OASS XML for Reis et al. (2008) contains a Pubmed ID for this Dias et al. (2007) paper [3]:

<pub-id pub-id-type=“pmid”>17992402</pub-id>

we can retrieve the PubMed bibliographic record for this paper by querying the Entrez API, from which we recover:

<Item Name=“AuthorList” Type=“List”>
<Item Name=“Author” Type=“String”>Dias JP</Item>
<Item Name=“Author” Type=“String”>Teixeira MG</Item>
<Item Name=“Author” Type=“String”>Costa MC</Item>
<Item Name=“Author” Type=“String”>Mendes CM</Item>
<Item Name=“Author” Type=“String”>Guimarães P</Item>
<Item Name=“Author” Type=“String”>Reis MG</Item>
<Item Name=“Author” Type=“String”>Ko A</Item>
<Item Name=“Author” Type=“String”>Barreto ML</Item>
<Item Name=“LastAuthor” Type=“String”>Barreto ML</Item>

Note that here we have the full author list containing the correctly accented surname “Guimarães”.

Thus, by matching this record with the original PLoS record for Dias et al., and selecting the longer list and the more accentuated names, we can correct the omissions in the original PubMed Central OASS data.

Example 4

However, Entrez output is not infallible. In the Entrez xml output for reference #44 in Reis et al. (2008), namely the paper by Travassos and Williams (2004) [4], we find:

<Item Name=“doi” Type=“String”>/S0102-311X2004000300003</Item>

The correct DOI for this article is doi:10.1590/S0102-311X2004000300003. However, the Entrez output from PubMed is missing the journal prefix “10.1590”, which when parsed to its URI form during our automated processing of the source data from XML to RDF would become


if we did not take steps to check for the correct DOI syntax.

These few examples, all taken from a single OASS article, illustrates some of the problems we have had to face in creating accurate and reliable RDF to enable us to publish these reference lists as open citation data.

What is shocking to me with regard to PLoS, perhaps the leading Open Access publisher, is that they don’t systematically include both DOIs and PubMed IDs in both HTML and XML versions of article references on the PLoS web site, despite the fact that they insert PubMed IDs into the records they mark up in NLM-DTD XML and send to PubMed Central, and also that PLoS persists with its policy of not listing all the authors, and that it does not include proper accents and diacritical marks, particularly for non-English names.

The methods used to correct citation errors are described in the next blog post, while the data processing pipeline through which we pass the input data to generate our RDF output Open Citations Corpus is described in the following blog post.

[1] Reis RB, Ribeiro GS, Felzemburgh RDM, Santana FS, Mohr S, Melendez SXTO, Queiroz A, Santos AC, Ravines RR, Tassinari WS, Carvalho MS, Reis MG, Ko AI
(2008). Impact of environment and social gradient on Leptospira infection in urban slums. PLoS Negl Trop Dis
2(4): e228. doi:10.1371/journal.pntd.0000228

[2]    Ko AI, Reis MG, Ribeiro Dourado CM, Johnson WD Jr, Riley LW and the Salvador Leptospirosis Study Group (1999). Urban epidemic of severe leptospirosis in Brazil. Lancet
354: 820–825. doi:10.1016/S0140-6736%2899%2980012-9.

[3]    Dias JP, Teixeira MG, Costa MC, Mendes CM, Guimarães P, Reis MG, Ko A, Barreto ML (2007). Factors associated with Leptospira sp infection in a large urban center in northeastern Brazil. Rev Soc Bras Med Trop. 40(5): 499-504. doi:10.1590/S0037-86822007000500002.

[4]     Travassos C, Williams DR (2004). The concept and measurement of race and their relationship to public health: A review focused on Brazil and the United States. Cad Saude Publica
20: 660–678. doi:10.1590/S0102-311X2004000300003.

Garbage in, garbage out – problems with bibliographic references

The Open Citations Project has aimed to liberate bibliographic references from biomedical research literature as Open Linked Data, using as its starting corpus the Open Access Subset (OASS) of articles within PubMed Central. The greatest problem faced during this project, naively unanticipated before we started, was the extend of incompleteness, noise and errors of various sorts within the reference information extracted from the OASS articles. So significant has this problem been, that it has taken almost the entire time and effort of Alex Dutton, our skilled data munger working on the project, to sort out, and without his skill, dedication and effort the project would not have succeeded.

In this context, any deviation of the bibliographic reference in an OASS article from the bibliographic citation text for that paper provided by the original publisher, is taken to be an error. These may be as slight as the substitution of a “β” character by the word “beta” in the title of a cited work by the author of the OASS article as he was creating his reference list, and as severe as including in the reference to one paper the DOI of another unrelated paper.

So it is worth taking time, before explaining how we addressed this problem, to describe its nature and magnitude, and to illustrate it with typical examples. We have found that errors of one sort or another occur in about 1% of all extracted references from the OASS. Since we extracted 6,325,178 individual references from our starting corpus, this constitutes well over 50,000 references containing errors.

Author errors

Errors have different sources. Authors are largely to blame, for not exercised due care when creating the reference lists of their paper (or earlier, when creating bibliographic records in a reference management system such as EndNote). However, if one of their EndNote records used to populate a reference list had been pulled automatically from some third-party source, the error might be due to that source, something of which the author was totally unaware.

Some references have incorrect punctuation or capitalisation of titles, or omit some sub-part, as exemplified by the following screenshot – note the use of “alternate”, “alternative” or “Alternate”, and of “intensivist” and “intensive” in different OASS references to the same paper. The boxed title is the correct one:

Some references omit one or more author names, or omit diacritics, a tendency particularly correlated with the degree to which a name is ‘foreign’ or ‘unusual’, as the next example illustrates – note the lack of “Mariotte-Labarre S” in the first two references, and the variation in punctuation of the journal name abbreviation in the third:

There are also numerous instances where the text of a reference is correct, but the associated identifiers (e.g. DOIs and PubMed IDs) are incorrect. By way of example, references 15 and 16 of PMC1839102 are both given the same DOI; in PMC2896208 the DOIs for references 52 and 72 are swapped; and in PMC2778786 references 15, 40, and 49 are all given the DOI of another (uncited) paper.

We should point out that these examples are entirely anecdotal, and that we haven’t investigated the frequency with which these or any other classes of error occurs.

Publisher errors

Publishers are the other main culprits, particularly for introducing errors into documents during the XML encoding stage, from which it is almost impossible to recover by the automated parsing systems we have used to extract information into our RDF-encoded records.

Individual publishers, while all working to the same National Library of Medicine DTD for encoding the XML markup of their articles submitted to PubMed Central, might take different approaches to encoding the same information. For example, the text

“… was found to be significant[1,3–6]

might be marked up as either of the following:

“… was found to be significant<sup>[<xref rid=”CR1″>1</xref>,<xref rid=”CR3″>3</xref>–<xref rid=”CR6″>6</xref>]</sup>”

“… was found to be significant<sup><xref rid=”CR1,CR3,CR4,CR5,CR6″>[1,3–6]</xref></sup>”.

Note that in the former case there is no explicit mention of the references with identifiers CR4 and CR5, making things a little harder to parse.

We have found occasions where a four-digit number in the title has been marked-up as a the publication year. For example, reference 21 of PMC2743650 claims that the cited article was published in 7942. The cited article’s real title refers to the bacterium strain PCC7942, information that has been removed from the title in the OACC reference.

The information returned by the Entrez API, which we used as our ‘gold standard’ against which to check OASS references containing PubMed IDs, was itself not without error. We found a number of PubMed records where DOIs had been truncated to just the prefix, or were missing a prefix.

Editors and referees

Others are also culpable. At the end of paragraph 12 of the PLoS ONE paper by Pickart et al. 2006 [1] we find the text:

We consider the current 12% detection rate to be a lower estimate of observable specific phenotypes from the screen, as additional screening will examine the morpholino collection using a variety of novel assays (such as newly generated enhancer and gene trap lines; Balciunas et al., 2004; Kawakami et al., 2004; Parinov et al., 2004) and may reveal developmental and/or functional aspects not readily visible by morphological criteria.

However, these three references do not appear in the reference list and so are totally lost to the system – the authors knew whom they were citing, but no-one else. How this escaped the eagle eyes of the authors, the journal editor and the reviewers is beyond my understanding!

In a separate blog post we describes how we have corrected some of these errors, while examples of errors in author lists are documented in the next post.

[1] Pickart MA et al. (2006). Genome-Wide Reverse Genetics Framework to Identify Novel Functions of the Vertebrate Secretome. PLoS ONE 1(1): e104. 10.1371/journal.pone.0000104.

Search OpenEdition Search

You will be redirected to OpenEdition Search