How to improve this code for getting pairwise? - r

Question

It is a question build upon the previous question (http://stackoverflow.com/questions/6538448/r-how-to-write-a-loop-to-get-a-matrix).
It is different from the previous one, as more details is provided, and libraries and example file is provided according to comments from DWin. So, I submitted it as a new question. Could you mind to teach me how to modify this code further?
To load the necessary libraries:
source("http://bioconductor.org/biocLite.R")
biocLite()
My protseq.fasta file has the following contents:
>drugbank_target|1 Peptidoglycan synthetase ftsI (DB00303)
MVKFNSSRKSGKSKKTIRKLTAPETVKQNKPQKVFEKCFMRGRYMLSTVLILLGLCALVARAAYVQSINADTLSNEADKR
SLRKDEVLSVRGSILDRNGQLLSVSVPMSAIVADPKTMLKENSLADKERIAALAEELGMTENDLVKKIEKNSKSGYLYLA
RQVELSKANYIRRLKIKGIILETEHRRFYPRVEEAAHVVGYTDIDGNGIEGIEKSFNSLLVGKDGSRTVRKDKRGNIVAH
ISDEKKYDAQDVTLSIDEKLQSMVYREIKKAVSENNAESGTAVLVDVRTGEVLAMATAPSYNPNNRVGVKSELMRNRAIT
DTFEPGSTVKPFVVLTALQRGVVKRDEIIDTTSFKLSGKEIVDVAPRAQQTLDEILMNSSNRGVSRLALRMPPSALMETY
QNAGLSKPTDLGLIGEQVGILNANRKRWADIERATVAYGYGITATPLQIARAYATLGSFGVYRPLSITKVDPPVIGKRVF
SEKITKDIVGILEKVAIKNKRAMVEGYRVGVKTGTARKIENGHYVNKYVAFTAGIAPISDPRYALVVLINDPKAGEYYGG
AVSAPVFSNIMGYALRANAIPQDAEAAENTTTKSAKRIVYIGEHKNQKVN
>drugbank_target|3 Histidine decarboxylase (DB00114; DB00117)
MMEPEEYRERGREMVDYICQYLSTVRERRVTPDVQPGYLRAQLPESAPEDPDSWDSIFGDIERIIMPGVVHWQSPHMHAY
YPALTSWPSLLGDMLADAINCLGFTWASSPACTELEMNVMDWLAKMLGLPEHFLHHHPSSQGGGVLQSTVSESTLIALLA
ARKNKILEMKTSEPDADESCLNARLVAYASDQAHSSVEKAGLISLVKMKFLPVDDNFSLRGEALQKAIEEDKQRGLVPVF
VCATLGTTGVCAFDCLSELGPICAREGLWLHIDAAYAGTAFLCPEFRGFLKGIEYADSFTFNPSKWMMVHFDCTGFWVKD
KYKLQQTFSVNPIYLRHANSGVATDFMHWQIPLSRRFRSVKLWFVIRSFGVKNLQAHVRHGTEMAKYFESLVRNDPSFEI
PAKRHLGLVVFRLKGPNCLTENVLKEIAKAGRLFLIPATIQDKLIIRFTVTSQFTTRDDILRDWNLIRDAATLILSQHCT
SQPSPRVGNLISQIRGARAWACGTSLQSVSGAGDDPVQARKIIKQPQRVGAGPMKRENGLHLETLLDPVDDCFSEEAPDA
TKHKLSSFLFSYLSVQTKKKTVRSLSCNSVPVSAQKPLPTEASVKNGGSSRVRIFSRFPEDMMMLKKSAFKKLIKFYSVP
SFPECSSQCGLQLPCCPLQAMV
>drugbank_target|5 Glutaminase liver isoform, mitochondrial (DB00130; DB00142)
MRSMKALQKALSRAGSHCGRGGWGHPSRSPLLGGGVRHHLSEAAAQGRETPHSHQPQHQDHDSSESGMLSRLGDLLFYTI
AEGQERTPIHKFTTALKATGLQTSDPRLRDCMSEMHRVVQESSSGGLLDRDLFRKCVSSSIVLLTQAFRKKFVIPDFEEF
TGHVDRIFEDVKELTGGKVAAYIPQLAKSNPDLWGVSLCTVDGQRHSVGHTKIPFCLQSCVKPLTYAISISTLGTDYVHK
FVGKEPSGLRYNKLSLDEEGIPHNPMVNAGAIVVSSLIKMDCNKAEKFDFVLQYLNKMAGNEYMGFSNATFQSEKETGDR
NYAIGYYHEEKKCFPKGVDMMAALDLYFQLCSVEVTCESGSVMAATLANGGICPITGESVLSAEAVRNTLSLMHSCGMYD
FSGQFAFHVGLPAKSAVSGAILLVVPNVMGMMCLSPPLDKLGNSHRGTSFCQKLVSLFNFHNYDNLRHCARKLDPRREGA
EIRNKTVVNLLFAAYSGDVSALRRFALSAMDMEQKDYDSRTALHVAAAEGHIEVVKFLIEACKVNPFAKDRWGNIPLDDA
VQFNHLEVVKLLQDYQDSYTLSETQAEAAAEALSKENLESMV
>drugbank_target|6 Coagulation factor XIII A chain (DB00130; DB01839; DB02340)
SETSRTAFGGRRAVPPNNSNAAEDDLPTVELQGVVPRGVNLQEFLNVTSVHLFKERWDTNKVDHHTDKYENNKLIVRRGQ
SFYVQIDFSRPYDPRRDLFRVEYVIGRYPQENKGTYIPVPIVSELQSGKWGAKIVMREDRSVRLSIQSSPKCIVGKFRMY
VAVWTPYGVLRTSRNPETDTYILFNPWCEDDAVYLDNEKEREEYVLNDIGVIFYGEVNDIKTRSWSYGQFEDGILDTCLY
VMDRAQMDLSGRGNPIKVSRVGSAMVNAKDDEGVLVGSWDNIYAYGVPPSAWTGSVDILLEYRSSENPVRYGQCWVFAGV
FNTFLRCLGIPARIVTNYFSAHDNDANLQMDIFLEEDGNVNSKLTKDSVWNYHCWNEAWMTRPDLPVGFGGWQAVDSTPQ
ENSDGMYRCGPASVQAIKHGHVCFQFDAPFVFAEVNSDLIYITAKKDGTHVVENVDATHIGKLIVTKQIGGDGMMDITDT
YKFQEGQEEERLALETALMYGAKKPLNTEGVMKSRSNVDMDFEVENAVLGKDFKLSITFRNNSHNRYTITAYLSANITFY
TGVPKAEFKKETFDVTLEPLSFKKEAVLIQAGEYMGQLLEQASLHFFVTARINETRDVLAKQKSTVLTIPEIIIKVRGTQ
VVGSDMTVTVQFTNPLKETLRNVWVHLDGPGVTRPMKKMFREIRPNSTVQWEEVCRPWVSGHRKLIASMSSDSLRHVYGE
LDVQIQRRPSM
To load the data to R for the analysis, I have done:
require("Biostrings")
data(BLOSUM100)
seqs <- readFASTA("./protseq.fasta", strip.descs=TRUE)
To get the the pairwise numbers, as there are a total of 4 sequences, I have done:
number <-c(1:4); dat <- expand.grid(number,number, stringsAsFactors=FALSE)
datr <- dat[dat[,1] > dat[,2] , ]
In order to calculate the score one by one, I can do this:
score(pairwiseAlignment(seqs[[x]]$seq, seqs[[y]]$seq, substitutionMatrix=BLOSUM100, gapOpening=0, gapExtension=-5))
However, I have problem to add a new column as "score" to include all the score for each pairs of the proteins. I tried to do this, but did not work.
datr$score <- lapply(datr, 1, function(i) { x <- datr[i,1]; y<- datr[i,2]; score(pairwiseAlignment(seqs[[x]]$seq, seqs[[y]]$seq, substitutionMatrix=BLOSUM100, gapOpening=0, gapExtension=-5))})
Could you mind to comments how to further improve it? Thanks DWin and diliop for wonderful solutions to my previous question.

Answer 1

Try:
datr$score <- sapply(1:nrow(datr), function(i) {
x <- datr[i,1]
y <- datr[i,2]
score(pairwiseAlignment(seqs[[x]]$seq, seqs[[y]]$seq, substitutionMatrix=BLOSUM100,gapOpening=0, gapExtension=-5))
})
To be able to reference your sequences better using their names, you might want to tidy up datr by doing the following:
colnames(datr) <- c("seq1id", "seq2id", "score")
datr$seq1name <- sapply(datr$seq1id, function(i) seqs[[i]]$desc)
datr$seq2name <- sapply(datr$seq2id, function(i) seqs[[i]]$desc)
Or if you just want to extract the accession IDs i.e. the contents of your parentheses, you could use stringr as such:
library(stringr)
datr$seq1name <- sapply(datr$seq2id, function(i) str_extract(seqs[[i]]$desc, "DB[0-9\\ ;DB]+"))
Hope this helps!