Timeline for Subset a protein fasta based on sequence length?
Current License: CC BY-SA 4.0
14 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 24, 2021 at 22:19 | vote | accept | R-MASHup | ||
| Aug 8, 2019 at 20:01 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Jul 11, 2019 at 8:24 | comment | added | terdon♦ |
For orthologs, I'd probably not pick since it's impossible to predict which protein will have ended up shared with the other organism. I'd just run a tblastn or better exonerate on using all proteins as queries. But yeah, picking the longest might make sense if you don't have a decent machine. Also, if my answer isn't working for you, you probably have ID lines that are different from what you've shown here. My solution just takes the 2nd field of the line ([gene=LOCNNNN] in your example), if that changes, it will not work.
|
|
| Jul 9, 2019 at 20:58 | comment | added | R-MASHup | Well it's a nonmodel organism, trying to find orthologs, and so need to reduce the number of entries somehow (esp as it's computationally so slow) ... how would you pick? | |
| Jul 9, 2019 at 20:47 | comment | added | terdon♦ | Yes. And choosing the longest is kind of arbitrary, but all this depends on what you're doing. For instance, I once needed to do something very similar because I was looking at various features of certain genes and I wanted to see if protein product length correlated with some other characteristic, so I chose the longest isoform of each gene's products. But, @R-MASHup, if you think the longest will be somehow special (most highly expressed or more representative or anything like that), you will be surprised. | |
| Jul 9, 2019 at 20:34 | vote | accept | R-MASHup | ||
| Jul 10, 2019 at 18:42 | |||||
| Jul 9, 2019 at 19:47 | comment | added | Daniel Standage | Perhaps more important is for R-MASHup to explain what is "strange" about these sequences and why. It is well known that a gene may encode multiple proteins, so the fact that there are multiple entries per gene doesn't surprise me. | |
| Jul 9, 2019 at 19:46 | answer | added | terdon♦ | timeline score: 1 | |
| Jul 9, 2019 at 19:30 | comment | added | terdon♦ |
NW_017095468 is a whole genome shotgun sequence, it will have many genes. The XP bits are RefSeq protein accessions. These can be different for different protein products of the same gene. It looks like the solutions will need to look at either the gene= or the GeneID: values.
|
|
| Jul 9, 2019 at 19:22 | comment | added | R-MASHup | I've added more of the file to my question. I think the unique/isoform bits are the XP_ (which has information about isoforms) ... NW_ bits are the genes. | |
| Jul 9, 2019 at 19:22 | history | edited | R-MASHup | CC BY-SA 4.0 |
added 3974 characters in body
|
| Jul 9, 2019 at 17:08 | comment | added | terdon♦ |
Can you give us an example that includes a duplicated gene? How is the "gene" defined? Should we look for the RefSeq protein ID (XP_123), for the gene=LOC string? All of the above? Will the entire header line be identical or only parts of it?
|
|
| Jul 9, 2019 at 16:40 | review | First posts | |||
| Jul 9, 2019 at 20:53 | |||||
| Jul 9, 2019 at 16:38 | history | asked | R-MASHup | CC BY-SA 4.0 |