I got an issue in SGVF_cmd.py

[Peihw]

In step 2，I run SGVFinder several times with the same data generated by Running ICRA，with the following command：

from SGVFinder2 import work_on_collection

vsgv, dsgv = work_on_collection( samp_to_map='example_out/', max_spacing=10, min_samp_cutoff=2, delsdetectthresh=0.25, real_del_thresh=0.95, dels_cooc_thresh=0.25, vsgv_dissim_thresh=0.125, vsgv_clip_quantile=0.02, vsgv_fit_interval=0.95, vsgv_fit_method='betaprime', x_coverage=0.01, rate_param=10, vsgv_dense_perc=85, browser_path=None, taxonomypath=DATABASE+'.taxonomy.df', genepospath=DATABASE+'genepos.df', frames_path=None ) vsgv.to_csv('vsgv_example.csv') dsgv.to_csv("dsgv_example.csv")

but each time， the output dsgv_example.csv was different from each other, with some lines randomly longer than the other lines. What may cause the formatting error?

[ym2877]

Hi @Peihw

To clarify- are the actual values different, or is the formatting just different?

[Peihw]

Hi @Peihw

To clarify- are the actual values different, or is the formatting just different?

I suppose it is formatting error. For the longer lines, the first few columns are exactly matched with the lines from other matrix. While the lengthen part with extra spaces are not easy to compare. I also found another issue that some column may turn into character in the output dsgv_example.csv.

[talkorem]

Can you please share the csv file with us? Either here, or find our e-mails at our lab website

[Peihw]

Can you please share the csv file with us? Either here, or find our e-mails at our lab website

I have sent the csv files to tal.korem@columbia.edu