Ranked Adjusted Rand: integrating distance and partition information in a measure of clustering agreement
<p>Abstract</p> <p>Background</p> <p>Biological information is commonly used to cluster or classify entities of interest such as genes, conditions, species or samples. However, different sources of data can be used to classify the same set of entities and methods allowi...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2007-02-01
|
| Series: | BMC Bioinformatics |
| Online Access: | http://www.biomedcentral.com/1471-2105/8/44 |
| _version_ | 1818790173238886400 |
|---|---|
| author | Carriço João A Pinto Francisco R Ramirez Mário Almeida Jonas S |
| author_facet | Carriço João A Pinto Francisco R Ramirez Mário Almeida Jonas S |
| author_sort | Carriço João A |
| collection | DOAJ |
| description | <p>Abstract</p> <p>Background</p> <p>Biological information is commonly used to cluster or classify entities of interest such as genes, conditions, species or samples. However, different sources of data can be used to classify the same set of entities and methods allowing the comparison of the performance of two data sources or the determination of how well a given classification agrees with another are frequently needed, especially in the absence of a universally accepted "gold standard" classification.</p> <p>Results</p> <p>Here, we describe a novel measure – the Ranked Adjusted Rand (<it>RAR</it>) index. <it>RAR </it>differs from existing methods by evaluating the extent of agreement between any two groupings, taking into account the intercluster distances. This characteristic is relevant to evaluate cases of pairs of entities grouped in the same cluster by one method and separated by another. The latter method may assign them to close neighbour clusters or, on the contrary, to clusters that are far apart from each other. <it>RAR </it>is applicable even when intercluster distance information is absent for both or one of the groupings. In the first case, <it>RAR </it>is equal to its predecessor, Adjusted Rand (<it>HA</it>) index. Artificially designed clusterings were used to demonstrate situations in which only <it>RAR </it>was able to detect differences in the grouping patterns. A study with larger simulated clusterings ensured that in realistic conditions, <it>RAR </it>is effectively integrating distance and partition information. The new method was applied to biological examples to compare 1) two microbial typing methods, 2) two gene regulatory network distances and 3) microarray gene expression data with pathway information. In the first application, one of the methods does not provide intercluster distances while the other originated a hierarchical clustering. <it>RAR </it>proved to be more sensitive than <it>HA </it>in the choice of a threshold for defining clusters in the hierarchical method that maximizes agreement between the results of both methods.</p> <p>Conclusion</p> <p><it>RAR </it>has its major advantage in combining cluster distance and partition information, while the previously available methods used only the latter. <it>RAR </it>should be used in the research problems were <it>HA </it>was previously used, because in the absence of inter cluster distance effects it is an equally effective measure, and in the presence of distance effects it is a more complete one.</p> |
| first_indexed | 2024-12-18T14:51:15Z |
| format | Article |
| id | doaj.art-adf2b61b29144333baa33c832f3ffc4c |
| institution | Directory Open Access Journal |
| issn | 1471-2105 |
| language | English |
| last_indexed | 2024-12-18T14:51:15Z |
| publishDate | 2007-02-01 |
| publisher | BMC |
| record_format | Article |
| series | BMC Bioinformatics |
| spelling | doaj.art-adf2b61b29144333baa33c832f3ffc4c2022-12-21T21:04:10ZengBMCBMC Bioinformatics1471-21052007-02-01814410.1186/1471-2105-8-44Ranked Adjusted Rand: integrating distance and partition information in a measure of clustering agreementCarriço João APinto Francisco RRamirez MárioAlmeida Jonas S<p>Abstract</p> <p>Background</p> <p>Biological information is commonly used to cluster or classify entities of interest such as genes, conditions, species or samples. However, different sources of data can be used to classify the same set of entities and methods allowing the comparison of the performance of two data sources or the determination of how well a given classification agrees with another are frequently needed, especially in the absence of a universally accepted "gold standard" classification.</p> <p>Results</p> <p>Here, we describe a novel measure – the Ranked Adjusted Rand (<it>RAR</it>) index. <it>RAR </it>differs from existing methods by evaluating the extent of agreement between any two groupings, taking into account the intercluster distances. This characteristic is relevant to evaluate cases of pairs of entities grouped in the same cluster by one method and separated by another. The latter method may assign them to close neighbour clusters or, on the contrary, to clusters that are far apart from each other. <it>RAR </it>is applicable even when intercluster distance information is absent for both or one of the groupings. In the first case, <it>RAR </it>is equal to its predecessor, Adjusted Rand (<it>HA</it>) index. Artificially designed clusterings were used to demonstrate situations in which only <it>RAR </it>was able to detect differences in the grouping patterns. A study with larger simulated clusterings ensured that in realistic conditions, <it>RAR </it>is effectively integrating distance and partition information. The new method was applied to biological examples to compare 1) two microbial typing methods, 2) two gene regulatory network distances and 3) microarray gene expression data with pathway information. In the first application, one of the methods does not provide intercluster distances while the other originated a hierarchical clustering. <it>RAR </it>proved to be more sensitive than <it>HA </it>in the choice of a threshold for defining clusters in the hierarchical method that maximizes agreement between the results of both methods.</p> <p>Conclusion</p> <p><it>RAR </it>has its major advantage in combining cluster distance and partition information, while the previously available methods used only the latter. <it>RAR </it>should be used in the research problems were <it>HA </it>was previously used, because in the absence of inter cluster distance effects it is an equally effective measure, and in the presence of distance effects it is a more complete one.</p>http://www.biomedcentral.com/1471-2105/8/44 |
| spellingShingle | Carriço João A Pinto Francisco R Ramirez Mário Almeida Jonas S Ranked Adjusted Rand: integrating distance and partition information in a measure of clustering agreement BMC Bioinformatics |
| title | Ranked Adjusted Rand: integrating distance and partition information in a measure of clustering agreement |
| title_full | Ranked Adjusted Rand: integrating distance and partition information in a measure of clustering agreement |
| title_fullStr | Ranked Adjusted Rand: integrating distance and partition information in a measure of clustering agreement |
| title_full_unstemmed | Ranked Adjusted Rand: integrating distance and partition information in a measure of clustering agreement |
| title_short | Ranked Adjusted Rand: integrating distance and partition information in a measure of clustering agreement |
| title_sort | ranked adjusted rand integrating distance and partition information in a measure of clustering agreement |
| url | http://www.biomedcentral.com/1471-2105/8/44 |
| work_keys_str_mv | AT carricojoaoa rankedadjustedrandintegratingdistanceandpartitioninformationinameasureofclusteringagreement AT pintofranciscor rankedadjustedrandintegratingdistanceandpartitioninformationinameasureofclusteringagreement AT ramirezmario rankedadjustedrandintegratingdistanceandpartitioninformationinameasureofclusteringagreement AT almeidajonass rankedadjustedrandintegratingdistanceandpartitioninformationinameasureofclusteringagreement |