Misc suggestions regarding "strand"

[cmdcolin]

1. possibility: add option not split out by strand. in a lot of experiments, the strand is not really important, it is basically random which strand you read from
2. possibility: investigate cases where strand is important. for example there is "strand-specific RNA-seq", a special wet-lab protocol, where the RNA is actually tagged properly with which strand it came from. this tells you whether it sequenced a positive or negative strand gene. see jbrowse option "first of pair strand"
3. investigate using SAM tags such as "XS" "TS" and "ts" instead of read strand to split out by strand. these tags are DIFFERENT from "sam flag read reverse strand" (https://broadinstitute.github.io/picard/explain-flags.html) and could be used to split out the a matrix differently from splitting out by the naive "sam flag reverse strand". they are sometimes related to detecting a canonical splice site. I can elaborate if interested.

see this jbrowse session showing a "strand-specific RNA-seq file" where the same data is colored in three ways: read strand (which is just a random mix of forward and reverse), "first of pair strand" (which basically matches the transcript it came from strand), and TS tag (also matches transcript it came from strand)

https://jbrowse.org/code/jb2/v2.17.0/?config=test_data%2Fconfig_demo.json&session=share-7yWY1Jgpux&password=vYddS

reverse strand gene:

forward strand gene

that there are groups who have already analyzed this (e.g. variation in single cell splicing)

[cmdcolin]

random brainstorming idea:

try applying algorithm to single cell rnaseq (scRNA-seq). single cell data makes it so each read is tagged with a unique "barcode" that corresponds to the cell it came from as i understand it.

this could reveal splicing patterns on a single cell level. this would create a matrix outputted for each cell probably, or make a series of matrixes or something like this

this is of course just my random idea, may or may not be in the cards for project goals.

that there are groups who have already analyzed this (e.g. variation in single cell splicing https://www.nature.com/articles/s41467-024-46480-9)

[cmdcolin]

xref #9

[cmdcolin]

you can see the effect of this in the bedgraph screenshot here

the negative vs positive split in the transition matrix is essentially due to randomness of the read strand rather than being biologically meaningful.

to recapitulate meaningful strand depends on data type (paired end stranded rna-seq requires flipping one of the pairs) or using accessory BAM tags (e.g. XS, TS, or ts)

[ihh]

Some experiments do have this info, as you say, so we probably want to keep this as an option (that by default collapses everything down to one direction)

…

On Mon, Nov 25, 2024 at 1:35 PM Colin Diesh ***@***.***> wrote: you can see the effect of this in the bedgraph screenshot here image.png (view on web) <https://github.com/user-attachments/assets/d1837abd-9909-4d58-91f3-872c641bf965> the negative vs positive split in the transition matrix is essentially due to randomness of the read strand rather than being biologically meaningful. to recapitulate meaningful strand depends on data type (paired end stranded rna-seq requires flipping one of the pairs) or using accessory BAM tags (e.g. XS, TS, or ts) — Reply to this email directly, view it on GitHub <#19 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAANGYPN7FQ72K7L3K5D5WT2COJ23AVCNFSM6AAAAABSHZWLSSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIOJZGA4DGOBSGI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>