Closed TS404 closed 6 years ago
Thank you.
Could you elaborate a bit on this?
Do you mean something like a tidy data frame:
scan_ag
- one row per P matched or one row per regex match?
From the perspective of plotting I trust it would be best if scan_ag had an output where each regex match was in one row with columns" id
, start
, end
.
predict_hyp
- a trimmed prediction
element where only predicted hydroxyprolines would be kept?
columns: id
, location
.
I could add an argument tidy = TRUE/ FALSE
to both functions which could provide such an output.
Yes, that's the sort of thing. I think that tidy dataframes are the way to go. Something similar to this output:
agregions <- scan_ag(sequence = sequences,
id = names(sequences),
dim = 3,
div = 6,
type = "extended", simplify = FALSE)$locations
agregions2 <- matrix(unlist(lapply(agregions,t)), ncol = 2, byrow = TRUE)
agregcount <- unlist(lapply(agregions, nrow))
agregions3 <- data.frame(y = rep(1:(length(sequences)),agregcount),
id = rep(names(sequences),agregcount),
align_start = agregions2[,1],
align_end = agregions2[,2])
scan_ag(sequence = at_nsp$sequence[c(1, 3, 16, 23)],
id = at_nsp$Transcript.id[c(1, 3, 16, 23)],
simplify = FALSE,
tidy = TRUE)[,-1] #to omit the sequence column from showing here
output:
id location.start location.end P_pos length AG_aa
1 ATCG00660.1 NA NA NA NA NA
2 AT2G28410.1 26 41 27, 35, 40 16 8
3 AT2G28410.1 55 70 55, 67, 69 16 6
4 AT2G43620.1 62 76 63, 70, .... 15 8
5 AT2G43620.1 167 185 168, 176.... 19 8
6 AT2G30933.1 36 51 37, 42, .... 16 9
7 AT2G30933.1 63 78 64, 68, .... 16 8
P_pos is a list column with P positions in the appropriate regex matches. length is the length of the matched substring. AG_aa is the number of amino acids in that were identified in the matched substring.
No information is lost compared to the list output.
What do you think?
Love it!
Added tidy
output. Currently it is available when simplify = FALSE
and tidy = TRUE
in function call.
The scan_ag and predict_hyp outputs are really nice.
It would also be good if there was an output option with the same colnames as "get_hmm" to simply list the locations of the relevant Prolines.