add sensitive mode

Expanded the comparison of N50 and gene counts to include details about the 20K genome set used for benchmarking. Clarified results regarding systematic bias and consistency of taxonomic labels between modes.

Author: yazhinia

sensitive.md

@@ -2,8 +2,8 @@
 
 Sensitive mode selects representative genomes based on connectivity, defined as the highest sum of weighted degrees in the ANI graph. It applies a greedy maximum weighted dominating set algorithm, where weighted degree of a bin is calculated as the sum of positive differences between the pairwise ANI and the ANI threshold.
 
-We compared N50 and gene counts between representative sets from sensitive and no-reassembly modes to assess whether representatives in sensitive mode exhibit systematic bias toward shorter sequences with inflated ANI to neighboring bins.
+We compared N50 and gene counts between representative sets from sensitive and no-reassembly modes to assess whether representatives in sensitive mode exhibit systematic bias toward shorter sequences with inflated ANI to neighboring bins. For this, we used 20K genome set that was used in benchmarking magmax and galah.
 
-Our results showed that there is no sytematic bias between two modes.
+Our results showed that there is no sytematic bias between two modes (upper and lower triangle values in figures show the percentage of representatives with higher N50 or gene counts relative to those selected by the other mode). Taxomic labels of the representatives and member genomes are consistent between both modes. 
 
 <img width="5319" height="2297" alt="n50_genecount_twomodecomp" src="https://github.com/user-attachments/assets/fe07ab10-d185-447e-b024-33c07a9d8e47" />