nunofonseca/msi

msi's Issues

R version question

For R, is 3.5+ required or does 4.1 work? The 4.1 trips the "ERROR: R version should be 3.5 or above" when the .1 is checked.

Conda: EMBOSS (rev) missing?

EMBOSS (revseq) is missing in the conda environment?

lca working well

Hi @nunofonseca

I compared the robitools lca with yours. Identical results. I tested 4 large vectors of between 1M and 4M rows. Using aggregate by qseqid$path or similar gives a single bin for each query (still identical to robitools). Yours is much faster! I will use it in my functions.

The only minor issue I have is that I start with 7 levels always and wish to end with 7 levels also (for other functions). To give exactly the same output I was getting before I need to follow your function with the code below. Perhaps this could be somehow built into the function?

lcaspNuno$binpath[is.na(lcaspNuno$binpath)]<-"unknown;unknown;unknown;unknown;unknown;unknown;unknown"
lcaspNuno$binpath[stringr::str_count(lcaspNuno$binpath,";")==5]<-paste0(lcaspNuno$binpath[stringr::str_count(lcaspNuno$binpath,";")==5],";unknown")
lcaspNuno$binpath[stringr::str_count(lcaspNuno$binpath,";")==4]<-paste0(lcaspNuno$binpath[stringr::str_count(lcaspNuno$binpath,";")==4],";unknown;unknown")
lcaspNuno$binpath[stringr::str_count(lcaspNuno$binpath,";")==3]<-paste0(lcaspNuno$binpath[stringr::str_count(lcaspNuno$binpath,";")==3],";unknown;unknown;unknown")
lcaspNuno$binpath[stringr::str_count(lcaspNuno$binpath,";")==2]<-paste0(lcaspNuno$binpath[stringr::str_count(lcaspNuno$binpath,";")==2],";unknown;unknown;unknown;unknown")
lcaspNuno$binpath[stringr::str_count(lcaspNuno$binpath,";")==1]<-paste0(lcaspNuno$binpath[stringr::str_count(lcaspNuno$binpath,";")==1],";unknown;unknown;unknown;unknown;unknown")
lcaspNuno$binpath[stringr::str_count(lcaspNuno$binpath,";")==0]<-paste0(lcaspNuno$binpath[stringr::str_count(lcaspNuno$binpath,";")==0],";unknown;unknown;unknown;unknown;unknown;unknown")

The only other thing I noticed is that using remove.dups is much slower the way I use the function (see below, I tried it a couple of times). btabsp here is a blast result of 1 million hits, from c. 21k queries.

> t1<-Sys.time()
> lcaspNuno = aggregate(btabsp$path, by=list(btabsp$qseqid),function(x) lca(x,sep=";",remove.dups = T))
> t2<-Sys.time()
> t3<-round(difftime(t2,t1,units = "mins"),digits = 2)
> t3
Time difference of 5.63 mins
> t1<-Sys.time()
> lcaspNuno = aggregate(btabsp$path, by=list(btabsp$qseqid),function(x) lca(x,sep=";"))
> t2<-Sys.time()
> t3<-round(difftime(t2,t1,units = "mins"),digits = 2)
> t3
Time difference of 2.13 mins
> 

zlib error

When running ./scripts/msi_install.sh -i ~/msi, during the execution of install_deps.sh, http://zlib.net/zlib-1.2.11.tar.gz is not found. The file has been moved to https://zlib.net/fossils/zlib-1.2.11.tar.gz.

Unclassified reads

MSI with Illumina paired end data

Hello,
Is it possible to use the MSI tool with Illumina paired-end data? If so, how should we proceed ?
Thank you.