Hi everyone, I am Anjali. I’m doing M.Sc Computational Biology currently in my last semester. I integrated AlphaGenome in my study on understanding the effect of ANRIL(lncRNA) variants on Cardiovascular diseases. It’s been 5 months playing around AlphaGenome and still I can’t understand and justify the results I have generated so far of the variant impact.
I also did Batch Variant Scoring but it seems like ontology terms for specific heart and brain tissue are missing and AlphaGenome scored variants across all the modalities, cell types, ontology and all.
I was so clueless recently because nobody around me understands AlphaGenome who could help me out.
I would appreciate any kind of help from this community and the scientists who have built this tool.
One thing to note is that AlphaGenome can not predict the effect of trans-regulation. I.e. it wouldn’t be able to predict how a different lncRNA sequence would have an effect on Cardiovascular disease. It can only predict the effects of variants on gene regulation, i.e. how would the expression or splicing of the gene change.
We don’t have all ontologies for all tissues. I would suggest using the closest available one.
I would also recommend first looking at score_variants output and see if there are any variants with quantile score of say 0.99 or more (especially for RNA-seq LFC and splice junctions) before looking at predict_variant plots.
1 Like
Right sir, this is one SNP’s plot. The rsID of this is - ‘rs10757278’ (known positive control for CVDs). This SNP is in ANRIL region and I was trying to study the effect of this SNP on ANRIL. ANRIL has two isoforms - linear and circular ANRIL. So what I was trying to find is - ‘Is this SNP anyhow contributes to the different ANRIL isoform formation and then correlate it with CVDs prediction’
I would really appreciate if you could help me understand this plot…this would definitely provide more clarity
