MetaChrom User Guide
Introduction
MetaChrom is a transfer learning framework that takes advantage of both an extensive compendium of publicly available chromatin profiles data, and epigenomic profiles of cell types related to specific phenotypes of interest. It's capable of predicting the genomic variant effect on epigenomic profiles with single-nucleotide resolution. Please see paper for details.
Input
Currently, MetaChrom accepts variant data from rsid or vcf files. Data can be either uploaded as .vcf or .txt files, or directly pasted to the text box.
VCF specifies the format of a text file for storing gene sequence variations. There should be precisely five columns in the VCF file, representing chromosome, position, id, reference allele, and alternative allele, respectively. The columns are separated by tabs. An example entry will be: chr1 109274968 test1 G T.
RSID stands for Reference SNP cluster ID. It is an accession number used by databases to refer to specific SNPs. MetaChrom accepts .txt files containing a list of RSIDs, where each RSID should be in a new line.
Output
MetaChrom outputs three regulatory feature predictions, only the absolute difference is shown in the result page, the other two can be downloaded in the result page.
Reference allele probabilities(Ref_prob): This is the predicted regulatory feature probablities for each cell type in our neurodevelopment model (please refer to our paper for details) for the reference allele.
Alternative allele probabilities(Alt_prob): This is the predicted regulatory feature probablities for each cell type in our neurodevelopment model (please refer to our paper for details) for the alternative allele.
Absolute Difference(Abs_diff): This is the absolute difference between the reference and alternative allele regulatory feature probabilities, which can be used to quantify the regulatory effect of the given variant. $${|Ref\_prob - Alt\_prob|}$$
Genome Coordinates
MetaChrom only supports GRCh38/hg38 genome coordinates.
Recommend Submmision Size
We recommend limiting the submission file size <10,000 entries. You will experience a long loading time when submitting a larger job. We suggest that you split up the large set into multiple smaller submissions or contact us directly.
