baudisgroup / segment-liftover Goto Github PK

Dear Prof. Gao
I appreciate your responses on previous issue. As my analysis going on, I found another issue that lead to the loss of large convertable segments. It primarily influence chr7 where large normal segments were recorded in our results(see data chunk). They all failed conversion from hg18 to hg19 probably caused by the start position chr7:52912 in hg18 is before the first mappable section(hg18 first map to hg19 is 102995, larger than 52912). These segments are mostly mappable by UCSC liftover that returns the hg19 start position as chr7:10238 (the start position of the first mappable section (chr7:10238) in hg19). I checked the UCSC genome browser in both genomic version and found the conversion is acceptable. So, I hope this feature can be supported in segment_liftover. Also, it may affect the end positions after the last section of a chromosome. This is hard to fix by myself. Please give me a hand. Thanks!

###################### DATA #########################
chr7 52912 6968498
chr7 52912 6536416
chr7 52912 158811019
chr7 52912 22401168
chr7 52912 52981165
chr15 19386005 20088817
chr7 52912 22401168
chr7 52912 22401168
chr15 19406785 20088817
chr7 52912 108398484
chr7 52912 57804817
chr7 52912 8287620
chr7 52912 158811019
chr7 52912 7950554
chr7 141896143 141908979
chr19 23793080 32457816
chr7 52912 6830773
chr7 52912 65204771
chr7 52912 158811019
chr7 52912 57670996
chr7 52912 10690406
chr7 52912 22401168
chr7 52912 4484678
chr7 52912 154024149
chr15 19537068 20013566
chr7 52912 5845530
chr7 52912 8278892
chr7 52912 3239052
chr7 52912 70059571
chr7 52912 22401168
chr7 52912 22401168
chr7 52912 7331134
chr7 142163568 142179254
chr7 52912 2556909
chr7 52912 61818476
chr7 52912 22401168
chr7 52912 141691434
chr7 52912 163647
chr7 52912 3366631
chr7 52912 22401168
chr17 22141081 22550352
chr7 52912 70058626
chr17 22185865 24648652
chr7 52912 38353316
chr7 52912 163647
chr17 22151867 22534030
chr7 52912 26108842
chr7 52912 8098531
chr7 141896143 141909358
chr7 52912 8525444
chr7 52912 76139971
chr7 52912 99376777
chr7 52912 878033
chr24 9253062 10006702
chr7 141699547 141706725
chr24 9253062 10006702
chr24 9253062 10006702
chr24 9253062 10036053
chr7 141699547 141715032
chr24 9253062 10122144
chr7 141699547 141720197
chr7 141720270 141725143
chr7 141699547 141706725
chr7 141699547 141720197
chr7 141699547 141712259
chr7 141699547 141706725
chr24 9253062 10006702
chr7 141699547 141712259
chr24 9253062 10006702
chr24 9253062 10006702
chr7 141699547 141706725
chr7 141707985 141891576
chr7 141699547 141715032
chr7 141699547 141706725

Dear Prof.
I found a bug when I run segment_liftover in batch mode. It just output one file which got the first file's name , but hold the last file's output. It is caused by a logic error in segmentLiftover.py near line 960 that the "if segment_output_file == None" statement only pass at first time. I hope you can fix this in furture version. Thanks.

J Yang

Hi @KyleGao ,

Is there a way to understand why a segment was not converted and if in some way a fine-tuning of a parameter may overcome this unconversion, please ?

For example an hg19 to hg38 conversion gave me :

chromosome start end same_chr/new_chr length_ratio/new_pos file
chr1 144854271 152052338 True 0.4237 input_dir/GE2486-Tumor_subclonal_allele_spe_cnv_0.43cellularity_2ploidy.transformed.txt
chr1 144621642 153916821 True 3.6042 input_dir/GE3344-Tumor_subclonal_allele_spe_cnv_0.55cellularity_2ploidy.transformed.txt

Thank you in advance for your help

Hi!
I have a still unpublished genome assembly, which I used to map and annotate my DNA-Seq pop data.
I want to use the ref genome for the zebra finch available from ensembl as my ancestral ref in some pop gen analysis. For this, I would have to convert the coordinates from the zebra finch available from ensembl to mine unpublished one.
Does this program fits to my problem?
Or would you recomend any other tool that would?
Thanks!

Dear segment-liftover team,
thank you for doing such a great job!
Our group is currently using your tool for making a comparison of orangutan, chimpanzee and human possible. Segment-liftover had already reduced the coordination loss drastically. Nonetheless, the humanized/lifted files of orangutan and chimpanzee are still containing peaks with their original chromosome names: chr2A and chr2B. These chromosomes are not part of the human genome and the peaks should be relocated somewhere on chr2. Do you have any ideas how I could solve the problem? I run the tool on default. Do you think that the problem is probably caused by the used chain files?

Kind regards
Sebastian

Some changes to the file should be implemented:

• consistency in headers & some logical structure (e.g. lower case; sample_id, probe_id, reference_name, segment_start, segment_end)
• I suggest making the "value" columns non-descriptive and open ended; i.e. value[1..n]. I would keep the value_1, value_2 ... example, but allow a undefined numer of columns (one could set a max. like 127, to block run-away data ...)
• I am not sure if we can include the liftover bibnary?
• I'm not happy with the "CNprobes" file name; so maybe just use "probes" (with an internal recognition of ^(?:CN)?probes(?:,chr\w\d?)\.tab$

Hi,
Thanks for making this amazing tool!
I ran the code
segment_liftover -i ./chm13 -o ./hg38 -c ~/references/t2t-chm13-v1.1.grch38.all.chain -si my.txt -so out.txt
and got the following error message.
object of type 'NoneType' has no len()
Could you please provide some guidance here? Thank you!

Cheers,
DK

As of pandas 2.0, pandas concat replaces append. Since the pandas version is not specified in the setup, the latest version is used as a default, which occasionally leads segment_liftover to fail.