https://github.com/functional-dark-side/functional-dark-side.github.io
Tip revision: 86968509e38902580b04a25786c5a58ba2777b21 authored by Antonio Fernandez-Guerra on 28 February 2022, 14:31:29 UTC
Merge pull request #6 from functional-dark-side/dependabot/bundler/nokogiri-1.13.3
Merge pull request #6 from functional-dark-side/dependabot/bundler/nokogiri-1.13.3
Tip revision: 8696850
4.2_Cluster_annotations.md
---
layout: page
title: Cluster annotations
---
The combination of clustering and functional annotation results led to two sets of clusters dominated by the not annotated one. As it is shown in figure 3(a), more than 60% (66.2%) (1,987,973 clusters) of the total ∼3 million clusters belong to clusters with not annotated members, while the clusters with annotated ORFs constitutes the 33% (1,015,924 clusters).
<img alt="cluster_annotations.png" src="/img/cluster_annotations.png" width="55%" height="" >
*Clusters subdivision based on functional annotations. a) Percentage of annotated and unannotated clusters. b) Annotated cluster size distribution. c) Not annotated cluster size distribution.*
However as illustrated in the second half of figure 3, the annotated clusters contain the 68% of the total ORFs (181,433,541 of 268,467,763) and have a larger average cluster size compared to the not annotated clusters as shown in figure 3(b) & 3(c).
- **Script and description:** The script [clu_annot.sh](scripts/Cluster_annotation/clu_annot.sh) combine the results of the clustering and the functional annotation. In a first step it retrieves the domain architecture (ordered multi-domain annotations) of each annotated ORFs (Pfam domain name, accession name, clan name), then it identifies two sets of clusters: annotated and not annotated, based on the presence of annotated ORFs in them.