generate RNA molecules from haploid genome

[yihangs]

Hi,

Thank you for developing such great software! I have one sample with pair-end reads (1.fastq, 2.fastq), and I want to simulate molecules based on this sample and the haploid reference genome without the need to phase diploid genomes. How can I achieve that?

I have the following observations:

1. Contradictory to the paper saying "BEAGLE requires at least two samples to perform its phasing operation", I use one sample as the input and camparee can still output two parental genomes.
2. The config file says that if we set the parameter "pooled" to be True, camparee will skip the phasing step. However, I tried it but still got two parental genomes.
3. I manually changed ploidy.txt file and set all the chromosomes to be haploid. The software failed at GenomeBuilderStep.

By the way, in the output file molecule_file.txt, each transcript id has the following form: 1_ENST00000564768.1_1/1_ENST00000564768.1_2, does "1_" represent the index of the sample, and "_1"/"_2" in the last represent the index of parental genome?

Also, there are some bugs in camparee_controller.py line 253, 257, "input_sample['filename']", "input_sample['filenames']". There is no parameter "filename" or "filenames" in our current config file.

Thanks!

[brainfood]

Hi,

Thank you so much for your interest in CAMPAREE and for finding those bugs. I'll include fixes for those in the next release. To broadly answer your question, CAMPAREE is designed to simulate diploid transcriptomes, which is why phasing and the generation of two parental genomes is baked into the pipeline. That being said, there are a couple of ways you can use CAMPAREE output or tweak its settings to effectively simulate data from a haploid genome.

1. Run CAMPAREE, but ignore the parental source information from the simulated transcripts in your downstream analyses. As you noted when looking at the simulated transcript IDs in the molecule_file.txt, the prefix identifies the sample the transcript originated from, and the suffix identifies the parental genome (your interpretation was exactly correct). If your downstream analyses ignore the parental genome information, it does not matter that they originated from a diploid simulation. This isn't a perfect solution because transcripts could have sequence differences between the two parental copies, so for some transcripts you'd end up with a mix of two different sequences for the same transcript. But, it means you can proceed immediately with any CAMPAREE data you've generated.

2. Modify the CAMPAREE output files to generate molecules from only one of the parental genomes. CAMPAREE lets you skip steps in the pipeline if you provide files that already contain the information generated by those steps. For example, if you already have aligned BAM files, you can use the config file to point to these BAM files and CAMPAREE will skip the genome alignment step. Similarly, if you've already run CAMPAREE once, you can use the output files created by that first run to skip all the steps in the pipeline and keep generating new molecules. Once you have the CAMPAREE output files in hand, you can find the allelic_imbalance_quantifications.txt file for each of the simulated samples. Its contents look like this:

#gene_id _1 _2
ENSMUSG00000000889 0.5 0.5
ENSMUSG00000001415 0.513 0.487
ENSMUSG00000001416 0.579 0.421
ENSMUSG00000001418 0.565 0.435
ENSMUSG00000001420 0.5 0.5
ENSMUSG00000001962 1.0 0.0
ENSMUSG00000001964 1.0 0.0
ENSMUSG00000002006 1.0 0.0
ENSMUSG00000002007 1.0 0.0

(continued from option 2.) This file defines what percentage of molecules generated from each gene originate from each parental genome/allele. In this example, half of the molecules simulated from ENSMUSG00000000889 will come from each parental allele. For ENSMUSG00000001962, all of the simulated molecules will come from a single parental allele (indexed '_1'). If you set all the values in the '_1' column to 1.0 and all of the values in the '_2' column to 0.0, you will effectively generate haploid data. Unlike with option 1, all of the simulated molecules will have the same sequences. You could also create this file directly, without running CAMPAREE first. You just need to make sure there is one entry for each gene ID in the annotation you're using. When you run CAMPAREE and provide it with this file for each sample (using the allele_quant: entries in the CAMPAREE config file), it will skip the execution of all the steps needed to calculate allelic imbalance, though it will still create the two parental genome sequences (even though the final molecules will only come from one of them).

Does that make sense? Please let me know if these options work for your needs. Thanks again for your interest.