-
Notifications
You must be signed in to change notification settings - Fork 11
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
hickit.js:457: Error #24
Comments
Hi Gang, It seems the BED file you're using (for Best, |
Hi Tan, Here is how I ran the program `./k8 hickit.js sam2seg -v b6_cast_snp_chr_parentheses.tsv aln.sam.gz | ./k8 hickit.js chronly - | ./k8 hickit.js bedflt par.bed - | gzip > contacts.seg.gz The sam and tsv files are too big to attach. Do I need to provide a bed file? I thought that's generated by the previous step. Thank you very much, Gang ` |
Hi Gang, This error is caused by missing the file Depending on the purpose of your analysis, a BED file If that is not your purpose, you can simply delete the step Best, |
Hi Tan. Thank you so much. I am kind of new in the field of data analysis. Did nor realize the need for the bed file. I do no have the file in hand. Can you please kindly provide it? Thanks Gang |
Hi Gang, Here I've attached the BED files for the human hg19 genome and for the mouse mm10 genome, respectively. The files are quite simple, containing just the (0-based) genomic coordinates of each PAR. hg19.par.bed.gz Best, |
Hi Tan, Thank you so much, Gang |
I got the information when after running, Does it mean there is something wrong with my SNP file? WARNING: a new allele A on read E00526:261:H773CCCX2:7:1101:11485:62980 at position chr4:37046608 (not C/T) |
This is completely normal. I typically redirect these warning messages to a log file, so they don't flood your screen: hickit -i contacts.seg.gz -o - 2> contacts.pairs.log | bgzip > contacts.pairs.gz |
Thank you, Tan, |
Can I skip the impute step? |
It looks the program does not allow me to use the raw contact.pair file directly.
|
Technically you can skip imputation; but you can only use contacts whose both legs are haplotype-resolved. In other words, for each (non-header) row of Therefore, to skip imputation, you must first remove any contacts with |
Than you Tan, I got my first reconstruction from my data. Best Gang |
Hi, I can't find the comment from your latest email, but you need to use the latest version of this repo to have the |
Hi Tan, Sorry that I delete the previous issue. Because I found that my .3dg does not have #unit line. I am regenerating the file. But it looks the new file does not have the line either. Everything looks fine except the #unit line. What may cause this? |
Again, you need to use a new version such as v0.1.1. |
I used |
is it 0.1 not 0.1.1? |
You downloaded an older release. |
Ok, got it. |
Hi Tan,@tanlongzhi I checked my imputation with Is it possible that there are too few hits with the threshold of 0.99? For example, two hits with a threshold of 0.99, one is correct but the other one is wrong. The following is how it looks: 0.99 0.5452 0.5000 0.4394 0.9999 0.4592 0.9999 |
Hi @chenggang108, I think your guess (low 50% imputation accuracy at 0.99 threshold is caused by statistical noise from small sample size) is probably right. If you'd like to look further, @lh3 once wrote an undocumented debug mode for imputation validation:
This mode should output true and imputed haplotypes. |
Hi @tanlongzhi Longzhi, Hope you are doing well. I am trying to figure out the calculation you did in the paper. Something I am not quite sure. I need to bother you again. My first question is about the unit used for all the values. I want to make sure that the units of the coordinate in the 3dg file is 'particle radii', right? So that the values from ''dip-c pd'' and ''dip-c dist'' are all with 'particle radii' as units. Am I right? The second question is about the potential somatic pairing. I am trying to figure out how you did it. Did you generate leg file for each haploid chromosome and did the pairwise distance? For example: Thanks |
Hi @chenggang108,
|
Thank you so much |
Hi @tanlongzhi , Sorry that I need to bother you again. I wonder is it true that we should get a very similar structure for one cell no matter how the seed is selected for hickit. For example, if I run:
I will have 100 3dg files for every resolution. If everything goes right, the RMSD between any two 3dg files from the 100 files should be very small, let's say RMSD< 3 or 2. The must be some wrong if it is not true. I am not sure whether I am right. Thanks Gang |
Yes, that's exactly how we determined which cell is good at which resolution. We took this RMSD concept (here implemented as "dip-c align") from Stevens et al. Nature 2017. Note that one must take the RMSD values with a grain of salt. The 3D modeling algorithm is not perfect, and may thus become overconfident at a structure (RMSD too low) or terrible at finding a good structure (RMSD too high). |
So, usually, RMSD must be < 2? |
I typically do RMSD ≤ 1.5. |
Thank you |
Hi @tanlongzhi, I have a question about the 3D-modeling of mitotic cells. Are there any prophase or metaphase cells among all the cells you analyzed? Does the program work on these cells? Thanks Gang |
We had one GM12878 cell at the very end of mitosis. It's a daughter cell right after cytokinesis. You can see its arrangement like one half of an anaphase cell. Please refer to the main figure of the paper for which cell. In general, mitotic cells stand out because of their very low inter-chromosomal contacts. I heard that Dr. Fraser's group managed to adapt the Dip-C algorithm for 3D-modeling mitotic cells (before cytokinesis); but I'm not aware of the details. |
Does the compaction of the genome affect the modeling? |
I don't think so. According to electron microscopy (e.g. ChromEMT), DNA density in mitotic cells is not that high. The main problem (before cytokinesis) is the presence of 4 copies (rather than 2) of each chromosome. |
Thank you |
Hi
I run into the problem as following, Can you please help with it?
Thanks
Gang
The text was updated successfully, but these errors were encountered: