Suppose there are two BioProjects in a tissue and they are not similar to each other and one of them has to be removed. Under these conditions, the current implementation of majority voting does not work well. In such cases, by referring to correlations with other tissues, it may be possible to pick out outliers that are not similar to any of the samples and exclude them. I'll find time to implement this, maybe in this or next week.
I haven't had time to work on this lately. I believe this is a minor improvement, so I'll leave it until someone is interested in or runs into a problem related to this.
Suppose there are two BioProjects in a tissue and they are not similar to each other and one of them has to be removed. Under these conditions, the current implementation of majority voting does not work well. In such cases, by referring to correlations with other tissues, it may be possible to pick out outliers that are not similar to any of the samples and exclude them. I'll find time to implement this, maybe in this or next week.