search

cozygene / FEAST

Fast expectation maximization for microbial source tracking
Other
116 stars 60 forks source link

Which the results in FEAST_output$data_prop can be used？ #29

Closed WeichuanLin6688 closed 2 years ago

WeichuanLin6688 commented 2 years ago

There is a difference between the two results （pred_emnoise_all and pred_em_all in FEAST_output$data_prop）. The pred_em_all result seems fit my experimental expectation. Which the results in FEAST_output$dataprop can be used？ ![287901646271759 pic](https://user-images.githubusercontent.com/73184954/156482625-762e83b3-ba67-4b0c-b83b-ab14d592d010.jpg)

liashenhav commented 2 years ago

I'm surprised there is such a discrepancy. Can you share your data files (feature table and metadata)? my email lis: liashenhav@gmail.com

Best, Liat

liashenhav commented 2 years ago

Thank you for sharing your data. The output of the algorithm is saved to a file with the name defined by the use (for example: outfile="Results"). This output is pred_emnoise_all.

timeresistance1996 commented 4 months ago

$data_prop pred_emnoise_all pred_em_all Env_2_ERR526061 0.1355451067 0.9288409664 Env_3_ERR526060 0.0584887804 0.0423748149 Env_4_ERR526063 0.0001485216 0.0005160769 unknown 0.8058175912 0.0282681418

When I tried the example data, I also encountered differences. What causes the huge difference between pred_emnoise_all and pred_em_all?Is the result of pred_emnoise_all more trustworthy?