Skip to main content

Table 1 Combinations of the dictionaries used in text mining to extract sentences showing relationships/co-mentions from public domain literature

From: A reverse metabolic approach to weaning: in silico identification of immune-beneficial infant gut bacteria, mining their metabolism for prebiotic feeds and sourcing these feeds in the natural product space

Relationship

No. of PubMed IDs

No. of Sent.

Co-mention

No. of PubMed IDs

No. of Sent.

gut-microbes | immune-related

89 (208)

172 (427)

gut-microbes | immune-related | infection-disease

42 (48)

66 (74)

gut-microbes | infection-disease

108 (316)

185 (564)

gut-microbes | immune-related | chemical-related

19 (74)

25 (132)

gut-microbes | chemical-related

0 (312)

0 (842)

gut-microbes | infection-disease | chemical-related

25 (58)

31 (75)

immune-related | infection-disease

0 (284)

0 (283)

immune-related | infection-disease | chemical-related

3 (7)

3 (49)

immune-related | chemical-related

0 (185)

0 (185)

gut-microbes | immune-related | infection-disease | chemical-related

2 (7)

2 (9)

infection-disease | chemical-related

0 (155)

0 (156)

   
  1. We constructed four dictionaries based on four concepts as described in the “Methods” section. The “Relationship” column shows the combinations of two dictionaries from the two concepts that were analyzed by natural language processing for extracting sentences containing a linguistic relation among both (Additional file 1). The “Co-mention” column shows the combinations of more than two concepts that were analyzed by natural language processing for extracting sentences mentioning those concepts (Additional file 2). For both “Relationship” and “Co-mention,” the table gives the number of PubMed IDs and the number of sentences that are extracted with the text mining pipeline after filtering for publication year (> 1999), human and mice studies, presence of microbial species, exclusion of pathogens and microbial genus names as described in the “Methods” section. Values before filtering are presented in parenthesis