Investigate better default value for --minCV
--minCV requires that the abundance of at least 1 library is greater than this value to be included in binning.
Having this value be the same as --minCVSum effectively means that --minCVSum is ignored. Additionally for libraries with many samples, the sum may be sufficiently high for confidence in the scafffold, but if no single sample has sufficiently high coverage to pass this filter the sequence would be completely ignored.
I believe that a lower number 0.1 - 0.33 may be more effective in capturing low abundance sequences in the binning process.
Additionally, it may be possible, to recruit scaffolds with a low but non-zero total mean coverage into existing clusters in the last phase where the lost and small contigs have been recruited. Or alternatively cluster just the low abundance scaffolds by TNF to themselves and possibly also the existing clusters which themselves are on the lower boundary of total abundance.
