SomaDataIO (version 6.2.0)

SeqId: Working with SomaLogic SeqIds

Description

The SeqId is the cornerstone used to uniquely identify SomaLogic analytes. SeqIds follow the format <Pool>-<Clone>_<Version>, for example "1234-56_7" can be represented as:

Pool	Clone	Version
1234	56	7

See Details below for the definition of each sub-unit. The <Pool>-<Clone> combination is sufficient to uniquely identify a specific analyte and therefore versions are no longer provided (though they may be present in legacy ADATs). The tools below enable users to extract, test, identify, compare, and manipulate SeqIds across assay runs and/or versions.

Usage

getSeqId(x, trim.version = FALSE)
regexSeqId()
locateSeqId(x, trailing = TRUE)
seqid2apt(x)
apt2seqid(x)
is.apt(x)
is.SeqId(x)
matchSeqIds(x, y, order.by.x = TRUE)
getSeqIdMatches(x, y, show = FALSE)

Value

getSeqId(): a character vector of SeqIds captured from a string.

regexSeqId(): a regular expression (regex) string pre-defined to match SomaLogic the SeqId pattern.

locateSeqId(): a data frame containing the start and stop

integer positions for SeqId matches at each value of x.

seqid2apt(): a character vector with the seq.* prefix, i.e. the inverse of getSeqId().

apt2seqid(): a character vector of SeqIds. is.SeqId() will return TRUE for all elements.

is.apt(), is.SeqId(): Logical. TRUE or FALSE.

matchSeqIds(): a character string corresponding to values in y of the intersect of x and y. If no matches are found, character(0).

getSeqIdMatches(): a \(n x 2\) data frame, where n is the length of the intersect of the matching SeqIds. The data frame is named by the passed arguments, x and y.

Arguments

x

Character. A vector of strings, usually analyte/feature column names, AptNames, or SeqIds. For seqid2apt(), a vector of SeqIds. For apt2seqid(), a character vector containing SeqIds. For matchSeqIds(), a vector of pattern matches containing SeqIds. Can be AptNames with GeneIDs, the seq.XXXX format, or even "naked" SeqIds.

trim.version

Logical. Whether to remove the version number, i.e. "1234-56_7" -> "1234-56". Primarily for legacy ADATs.

trailing

Logical. Should the regular expression explicitly specify trailing SeqId pattern match, i.e. "regex$"? This is the most common case and the default.

y

Character. A second vector of AptNames containing SeqIds to match against those in contained in x. For matchSeqIds() these values are returned if there are matching elements.

order.by.x

Logical. Order the returned character string by the x (first) argument?

show

Logical. Return the data frame visibly?

Functions

• getSeqId(): extracts/captures the the SeqId match from an analyte column identifier, i.e. column name of an ADAT loaded with read_adat(). Assumes the SeqId pattern occurs at the end of the string, which for the vast majority of cases will be true. For edge cases, see the trailing argument to locateSeqId().

• regexSeqId(): generates a pre-formatted regular expression for matching of SeqIds. Note the trailing match, which is most commonly required, but locateSeqId() offers an alternative to mach anywhere in a string. Used internally in many utility functions

• locateSeqId(): generates a data frame of the positional SeqId matches. Specifically designed to facilitate SeqId extraction via substr(). Similar to stringr::str_locate().

• seqid2apt(): converts a SeqId into anonymous-AptName format, i.e. 1234-56 -> seq.1234.56. Version numbers (1234-56_ver) are always trimmed when present.

• apt2seqid(): converts an anonymous-AptName into SeqId format, i.e. seq.1234.56 -> 1234-56. Version numbers (seq.1234.56.ver) are always trimmed when present.

• is.apt(): regular expression match to determine if a string contains a SeqId, and thus is probably an AptName format string. Both legacy EntrezGeneSymbol-SeqId combinations or newer so-called "anonymous-AptNames" formats (seq.1234.45) are matched.

• is.SeqId(): tests for SeqId format, i.e. values returned from getSeqId() will always return TRUE.

• matchSeqIds(): matches two character vectors on the basis of their intersecting SeqIds. Note that elements in y not containing a SeqId regular expression are silently dropped.

• getSeqIdMatches(): matches two character vectors on the basis of their intersecting SeqIds only (irrespective of the GeneID-prefix). This produces a two-column data frame which then can be used as to map between the two sets.

  The final order of the matches/rows is by the input corresponding to the first argument (x).

  By default the data frame is invisibly returned to avoid dumping excess output to the console (see the show = argument.)

Author

Stu Field

Details

Pool:	ties back to the original well during SELEX
Clone:	ties to the specific sequence within a pool
Version:	refers to custom modifications (optional/defunct)

AptName

a SeqId combined with a string, usually a GeneId- or seq.-prefix, for convenient, human-readable manipulation from within R.

See Also

intersect()

Examples

x <- c("ABDC.3948.48.2", "3948.88",
       "3948.48.2", "3948-48_2", "3948.48.2",
       "3948-48_2", "3948-88",
       "My.Favorite.Apt.3948.88.9")

tibble::tibble(orig       = x,
               SeqId      = getSeqId(x),
               SeqId_trim = getSeqId(x, TRUE),
               AptName    = seqid2apt(SeqId))

# Logical Matching
is.apt("AGR2.4959.2") # TRUE
is.apt("seq.4959.2")  # TRUE
is.apt("4959-2")      # TRUE
is.apt("AGR2")        # FALSE


# SeqId Matching
x <- c("seq.4554.56", "seq.3714.49", "PlateId")
y <- c("Group", "3714-49", "Assay", "4554-56")
matchSeqIds(x, y)
matchSeqIds(x, y, order.by.x = FALSE)

# vector of features
feats <- getAnalytes(example_data)

match_df <- getSeqIdMatches(feats[1:100], feats[90:500])  # 11 overlapping
match_df

a <- utils::head(feats, 15)
b <- withr::with_seed(99, sample(getSeqId(a)))   # => SeqId & shuffle
(getSeqIdMatches(a, b))                          # sorted by first vector "a"