Take the 2-minute tour ×
Biology Stack Exchange is a question and answer site for biology researchers, academics, and students. It's 100% free, no registration required.

I am trying to convert a set of Affymetrix ID's, like this one 143053_at_3745, to Flybase Gene Numbers (FBgn) like this one FBgn0000015. I have downloaded the Flybase file required to do so (as described here) but I have noticed that most of the FBgn's have more than one associated Affy id.

My question is, how do I know which to assign to my data? From the list of affy-ids that I have, how do I label each one with the appropriate FBgn? I am using data from the Ayroles et al 2009 (DGRP) which has a column of AffyID which look similar to the ones in the Flybase file but shorter (1638273_at). Perhaps I don't grasp why there would be more than one.

Some have one affy id which is repeated across 15 columns in the file, whereas some seem to have several affy id's associated with them.

Some questions this problem raises:

Why does a singular FBgn identifier have more than one associated affy identifier?

Why do some datasets, such as the dataset I have, have shorter versions of the affy identifier?

Where can I find an up to date and appropriate list matching these Affy id's the FBgn identifiers correctly?

share|improve this question
    
perhaps @shigeta knows this... –  GriffinEvo Mar 25 '13 at 8:32
    
Is it possible that there are multiple probes for the same gene? You've said that most of the FBgns have multiple Affy IDs associated with them, but you haven't said the converse. –  dd3 Mar 25 '13 at 15:25
    
I suspect that is the case, but I am not familiar with the technicalities of affymetrix chip design, and was confused as to why there would be some with the same - though further examination shows that the latter part (after the final underscore) changes within that subset. –  GriffinEvo Mar 25 '13 at 16:27
add comment

1 Answer

up vote 3 down vote accepted

I can answer this - I may not have time to dig though the file you are pointing to... but here's some explanation - lmk if you need more.

The shorter names (123456_at) are the original names for the probe sets that Affymetrix gave. The file you area asking about has been extended for FlyBase's purpose and I'm only dimly aware of its existence. It looks like Flybase has tried to rename the probe set to a minimal gene list and create less ambiguous mappings. I'm not familiar with it. If I have time Ill try to look at it and put something here, but I'm fairly slammed this week.

In general you should know, there is a many to many relationship between probe sets from arrays and genes.

There are a few reasons for this. The most common reasons:

More than one probe set per gene 1) More than one start sequence per gene. The IVT arrays such as you are looking at read only the 3' end of the gene. If there are more than one such terminii, you will have more than one probe set. 2) Duplicate genes. If the 3' end of the gene has been duplicated recently, then a probe set may read more than one gene and not be able to distinguish them, so it will have both gene references in its annotation files. 3) Duplicate probe sets. For older arrays, this did happen where the probes in two probe sets will be mostly or entirely the same. 4) formerly separate genes are now joined into a single gene. This is similar to (1) above, but with the added reason that two neighboring transcripts seen separately at the time of the array design are currently known to be part of the same transcript. This is something we often don't think about, but the array may have been designed before your current gene of interest had a full length sequence.

In several cases the transcription behavior of the two probe sets will be quite different and you can decide which one to take based on how it behaves experimentally. for example one of the probe sets may never respond while another one may register large changes with biological sample conditions. Sorry I can't be of more help here. There are too many possible scenarios to consider to write about cogently.

More than one gene per probe set 1) a single stretch of DNA, even on the same strand, may simply be associated with more than one gene and the probe set will read both of them. 2) Even if they are not exactly the same, some genes resemble each other in nucleotide sequence that is impossible to choose probes which do not read both of them.

Since they have similar read length to a probe set, all this is also true to some extent for short read NGS sequencers in RNASeq data.

As you can hopefully see, ambiguity is something you have to expect to some extent when deciding to work with a biological system. In many cases (~ 80%) you will have a single gene read per probe set, but with those odds you'll be looking at the genome browser for a half dozen of your favorite experimental results.

share|improve this answer
1  
excellent explanation, thanks very much! –  GriffinEvo Mar 25 '13 at 16:57
    
glad it sufficed - there's always more... lol –  shigeta Mar 25 '13 at 17:56
    
seemingly so... –  GriffinEvo Mar 25 '13 at 18:22
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.