Death case numbers were compared between clusters with high-risk mortality and those lying outside the clusters.
A hierarchical cluster dendrogram was generated using Yule's Q as the similarity measure between clusters, with a higher value indicating greater similarity measurement.
This may be the result of large overlaps between gene clusters with high transcription levels at different stages, as a large proportion of genes has high transcriptional activity throughout the erythrocytic cell cycle.
Hierarchical Agglomerative Clustering with Overlap (HACO) [ 4] first considers all vertices as individual clusters, then iteratively merges pairs of clusters with high connectivity between them.
While all clusters with high correlation between field environments (r > 0.8) had ED models that fitted the cluster mean well (model MSE < 0.12), low MSE models were also selected for several clusters with strong dissimilarities between field environments.
Others use the ratio of within-cluster similarity to between-cluster similarity, with higher ratios denoting a better cluster solution [45], often times using the same feature upon which the clustering was based.
The main difference between ORFans and non-ORFans is an excess of clusters with high d N values.
These kinds of evaluation usually assign the best score to the algorithm that produces clusters with high similarity within a cluster and low similarity between clusters.
A good clustering algorithm must produce high-quality clusters with high levels of intra-cluster similarity and with low levels of inter-cluster similarity.
Clusters with high or low HCV prevalence, and clusters with high or low PAT exposure, were scattered, with no evident overall pattern of an association between HCV prevalence and PAT exposure.
