search

immunoengineer / gliph

Grouping of Lymphocyte Interactions by Paratope Hotspots
GNU General Public License v3.0
90 stars 30 forks source link

Cluster #3

Open weizhiting opened 6 years ago

weizhiting commented 6 years ago

Hi, use your perl script and second validation data, GLIPH only cluster 4% of all TCRs, i use all the default parameter, am i missing something? Thanks!

TracyYXChen commented 6 years ago

Same here! In the M. tuberculosis data, I also only clustered ~4% TCRs, but the number is pretty close, I got 41 clusters with a size >= 3. However, I checked the intersection between my clusters and those 43 clusters in Supplementary Information S6 and found we only have about 1/3 same clusters...

weixuanfu commented 4 years ago

Is there a detailed documentation to reproduce the results in the paper?