mccannj9 / ace-kit Goto Github PK

Almitey is run on each cluster of the RepeatExplorer output. It would be good to have an HTML table which summarizes and has links to the results HTML tables from the individual almitey runs on each cluster (similar to the table that is output for the general RepeatExplorer outputs).

python start.py -a /mnt/raid/users/tmp/rypos/seqclust/clustering/clusters/dir_CL0014/contigs.ace -o tmp`

CL14Contig48
CL14Contig79
CL14Contig173
CL14Contig334
CL14Contig432
Traceback (most recent call last):
File "start.py", line 52, in
results_dict, all_reads = finder.fit()
File "/mnt/raid/users/petr/workspace/ace-kit/kit/finder.py", line 33, in fit
self.generate_output_for_boundaries(cands, derivs)
File "/mnt/raid/users/petr/workspace/ace-kit/kit/finder.py", line 48, in generate_output_for_boundaries
boundary.set_logo(save=logo_out)
File "/mnt/raid/users/petr/workspace/ace-kit/kit/contig.py", line 315, in set_logo
df = create_seqlogo_dataframe(seqs)
File "/mnt/raid/users/petr/workspace/ace-kit/kit/utils.py", line 58, in create_seqlogo_dataframe
seqlen = len(sequences[0])
IndexError: list index out of range

I had to add txt suffix to ace file to be able to upload it to github:

contigs.ace.txt

First, boundary sequence and sequence logos are shifted by one relative to each other. Second, adding more of the outer surrounding region from the boundary to see more visually the low information regions would be useful. Apply information threshold to only the one side of the sequence logo corresponding to the side of the insertion site.

so it can be is as function:

if name == 'main':
almitey(...)

in init default value for logfile is none - which is OK latet it does not work with this default.
I suggest it is set to input_dir/log.txt if logfile is none

both pandas and logomaker are available in bioconda channel

Shorter contigs (<1 kb) and 2 insertion site boundaries may be indicative of a MITE or SINE element. These statistics can be used as a possible "pre"-annotation to indicate to users that a particular cluster may be a non-autonomous transposable element.

in init default value for logfile is none - which is OK latet it does not work with this default.
I suggest it is set to input_dir/log.txt if logfile is none

Currently, the sequence logo construction function only uses reads which directly overlap the insertion site boundaries. Ideally, it should also use information from reads which do not overlap the boundary, but still overlap the region surrounding the boundary that is taken into account for computing the sequence logo dataframes.