Journal: Molecular Cell

Article Title: A CRISPR Resource for Individual, Combinatorial, or Multiplexed Gene Knockout

doi: 10.1016/j.molcel.2017.06.030

Figure Lengend Snippet: CRoatan, an Algorithm for Identifying Potent sgRNAs (A) The potency of sgRNAs analyzed in Doench et al. stratified by conservation score (calculated as described in Figure S1 B, ρ = 0.32). An sgRNA percentile is the percentile rank of an sgRNA relative to all other effectors targeting the same gene. This plot, as all others in the figure, was generated with the MATLAB boxplot function using default parameters. The edges of the box are the 25 th and 75 th percentiles. The error bars extend to the values q3 + w(q3 − q1) and q1 − w(q3 − q1), where w is 1.5 and q1 and q3 are the 25 th and 75 th percentiles. (B) Efficacy percentiles of the sgRNAs analyzed in Doench et al. when stratified by the likelihood of frameshift mutations (FSM likelihood) at the corresponding target site (rank-sum p value = 0.0405 for tertile 3 versus tertile 1 and 2 sgRNAs). (C) Efficacy percentiles of the sgRNAs analyzed in Doench et al. when stratified by the consolidated CRoatan algorithm (ρ = 0.52). (D) Z -score-normalized depletion rates of EG-sgRNAs when stratified by CRoatan score (ρ = 0.21). Depletion rates were calculated as the average log ratio in screens carried out in A-375 and K-562 cells. (E) Depletion rates of NEG- and EG-targeting sgRNAs in screens corresponding to those described in (D). sgRNA libraries were designed using the GPP-WP, sgRNAScorer, and Edit-R algorithms (rank-sum p value = 0.0942 for GPP-WP, 0.0209 for sgRNAScorer, and 0.0233 for Edit-R).

Article Snippet: The 10 forests were trained using the MATLAB treeBagger package (1000 trees per forest).

Techniques: Generated

Journal: Molecular Cell

Article Title: A CRISPR Resource for Individual, Combinatorial, or Multiplexed Gene Knockout

doi: 10.1016/j.molcel.2017.06.030

Figure Lengend Snippet: Dual-CRoatan Constructs Provide Superior CRISPR-Based Gene Targeting (A) Average depletion rates for each EG-sgRNA when it is paired with NEG-sgRNAs (gray) and when it is paired with sgRNAs targeting the same EG (brown). sgRNAs are grouped based on the gene target; rank-sum p value = 0.0006. (B) Depletion rates of NEG- and EG-targeting CRISPR constructs in negative-selection screens. Shown are the consolidated depletion rates for single-sgRNA constructs selected using pre-existing tools (GPP, sgRNAScorer, or Edit-R algorithms) as well as the rates for CRoatan single-sgRNA constructs and CRoatan dual-sgRNA constructs (dual-CRoatan, rank-sum p value = 2.5e -5 for existing algorithms and p value > 0.05 for single-CRoatan constructs). Depletion rates were calculated as the average log ratio in screens carried out in A-375 and K-562 cells. This plot was generated with the MATLAB boxplot function using default parameters. The edges of the box are the 25 th and 75 th percentiles. The error bars extend to the values q3 + w(q3 − q1) and q1 − w(q3 − q1), where w is 1.5 and q1 and q3 are the 25 th and 75 th percentiles. (C) Gene-level analysis of CRoatan and CRoatan dual-sgRNA construct depletion rates. Using the average depletion rate for each construct in A-375 and K-562 cells, gene “hits” were calculated using a series of stringencies (top 10%, 20%, 30%, 40%, and 50% most-depleted sgRNAs). For a gene to be called a hit at a given stringency, a minimum of two constructs need to be depleted beyond the stringency level.

Article Snippet: The 10 forests were trained using the MATLAB treeBagger package (1000 trees per forest).

Techniques: Construct, CRISPR, Selection, Generated